Description
This paper analyses the determinants of broadband Internet access prices
in a group of 15 EU countries between 2008 and 2011. Using a rich panel
dataset of broadband plans, we show the positive effect of downstream
speed on prices, and report that cable and fibre-to-the-home technologies
are available at lower prices per Mbps than xDSL technology. Operators’
marketing strategies are also analysed as we show how much prices rise
when the broadband service is offered in a bundle with voice telephony
and/or television, and how much they fall when download volume caps
are included. The most insightful results of this study are provided by a
group of metrics that represent the situation of competition and entry
patterns in the broadband market. We show that consumer segmentation
positively affects prices. On the other hand, broadband prices are higher
in countries where entrants make greater use of bitstream access and lower
when they use more intensively direct access (local loop unbundling).
However, we do not find a significant effect of inter-platform competition
on prices.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2011/07 pàg. 1
Research Institute of Applied Economics Working Paper 2011/07 pag .1
1
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 33 pàg.
Research Institute of Applied Economics Working Paper 2013/09 33 pag.
“Broadband prices in the European Union: competition and
commercial strategies”
Joan Calzada and Fernando Martínez
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 2
Research Institute of Applied Economics Working Paper 2013/09 pag. 2
2
WEBSITE: www.ub.edu/irea/ • CONTACT: [email protected]
The Research Institute of Applied Economics (IREA) in Barcelona was founded in 2005, as a
research institute in applied economics. Three consolidated research groups make up the
institute: AQR, RISK and GiM, and a large number of members are involved in the Institute. IREA
focuses on four priority lines of investigation: (i) the quantitative study of regional and urban
economic activity and analysis of regional and local economic policies, (ii) study of public
economic activity in markets, particularly in the fields of empirical evaluation of privatization, the
regulation and competition in the markets of public services using state of industrial economy, (iii)
risk analysis in finance and insurance, and (iv) the development of micro and macro econometrics
applied for the analysis of economic activity, particularly for quantitative evaluation of public
policies.
IREA Working Papers often represent preliminary work and are circulated to encourage
discussion. Citation of such a paper should account for its provisional character. For that reason,
IREA Working Papers may not be reproduced or distributed without the written consent of the
author. A revised version may be available directly from the author.
Any opinions expressed here are those of the author(s) and not those of IREA. Research
published in this series may include views on policy, but the institute itself takes no institutional
policy positions.
1
Broadband prices in the European Union:
competition and commercial strategies
#
Joan Calzada
*
and Fernando Martínez
+
May 2013
Abstract
This paper analyses the determinants of broadband Internet access prices
in a group of 15 EU countries between 2008 and 2011. Using a rich panel
dataset of broadband plans, we show the positive effect of downstream
speed on prices, and report that cable and fibre-to-the-home technologies
are available at lower prices per Mbps than xDSL technology. Operators’
marketing strategies are also analysed as we show how much prices rise
when the broadband service is offered in a bundle with voice telephony
and/or television, and how much they fall when download volume caps
are included. The most insightful results of this study are provided by a
group of metrics that represent the situation of competition and entry
patterns in the broadband market. We show that consumer segmentation
positively affects prices. On the other hand, broadband prices are higher
in countries where entrants make greater use of bitstream access and lower
when they use more intensively direct access (local loop unbundling).
However, we do not find a significant effect of inter-platform competition
on prices.
Keywords: Telecommunications, Broadband prices, European Union,
Competition, Regulation
JEL: L51, L86, L96.
#
We thank Joan Ramon Borrell, Martin Cave, Xavier Fageda, Michal Grajek, Iñigo Herguera, Daniel
Montolio, Qmars Safikhani, Tommaso Valletti and Carlos Velasco for their useful comments and
suggestions. We also thank seminar participants at CMT (Barcelona), ITS (Viena), JEI (Murcia), Senatel
(Quito) and the Workshop on Telecommunications Economics (Barcelona). This research has received
financial support from the Spanish Ministry of Economics and Competitiveness (ECO2012-33392) and the
Regional Government of Catalonia (SGR2009-1066).
*
Corresponding author: Joan Calzada. Departament de Política Econòmica. Universitat de Barcelona.
Avinguda Diagonal 690. 08520 Barcelona. Tel. +34 934 034 743. Email: [email protected].
+
Fernando Martínez: Competition Commission and Universitat de Barcelona. Departament de Política
Econòmica. Avinguda Diagonal 690. 08520 Barcelona. Email: [email protected]
2
1. Introduction
Over the last decade, millions of people in the European Union have installed broadband
in their households
1
, enabling them to download information and use sophisticated digital
services
2
. Broadband Internet access is an essential component of inclusiveness in the 21
st
century, and households without access to this service are in risk of becoming marginalized
from society and economic opportunity. Several papers have analyzed the impact that
technological change and regulation have had in the expansion of Internet. However, little
attention has been given to how telecommunication operators have modified their business
strategies to respond to market evolution and competition. A detailed analysis of the
managerial behaviour of operators is essential if we are to understand the significant price
and quality differences that have emerged across Europe.
This paper examines the factors determining fixed broadband Internet prices in 15 EU
countries between 2008 and 2011
3
. We employ a rich data set that contains both the
commercial and technical characteristics of 2,204 plans offered by incumbent and entrant
operators. By using an instrumental variable approach we estimate a pricing equation using
three types of variable: (1) the technical characteristics of the plans; (2) the operators’
commercial strategies; and (3) the patterns of competition in the country. To the best of
our knowledge, this is the first paper to use information at the level of the individual firm’s
commercial plans to examine the determinants of prices in the EU.
Our study shows how operators adjust their prices to the technological characteristics of
the plans. First, we show that downstream speed has a positive and significant non-linear
impact on price: a 10 Mbps increase in the downstream speed increases the price by around
5%. And second, after controlling for downstream speed, we find that cable or fibre-to-
the-home (FTTH) broadband plans are cheaper than xDSL offers. This result is important
for orienting regulatory policy in the sector, as it questions the interest that operators might
have for deploying Next Generation Access Networks (NGAs).
This paper also shows the importance of the operators’ commercial practices. Thus while
flat rate plans are more expensive than metered plans (which are limited in volume), plans
that offer a broadband service bundled with voice telephony and/or television are
considerably higher in price, especially in the case of triple packages. Given the importance
of such bundling practices in many European markets today, we also analyse the main
determinants of this commercial practice.
Finally, we examine how levels of competition and regulation affect operators’ pricing
decisions. We show that incumbents set prices that are often significantly higher than those
of their competitors, which might be a consequence of such factors as their wider
1
The European Commission (2008) defines broadband internet access as “an access assuring an always-on
service with speeds in excess of 144 kbps. This speed is measured in download terms.”
2
During the nineties, broadband was delivered over cable and telephone lines. In the years that followed,
these technologies where upgraded and some operators began to deploy fibre to the home delivery, which
supports higher bandwidth (Greenstein and McDevitt, 2011a and 2011b).
3
In spite of their growing relevance, broadband mobile services are not included in this analysis. Note that
the commercial characteristics of mobile offers differ markedly from those of fixed broadband Internet
access. For example, download speed is significantly slower in the case of mobile offers (although new
wireless technologies such as LTE can provide speeds similar to fixed broadband)
3
coverage, reputation, or the existence of consumer switching costs. In contrast, the degree
of price dispersion and the overall number of plans offered in any given country have a
positive effect on prices. This result illustrates that market segmentation and consumer
confusion in the face of various marketing plans allow operators to increase their prices.
A further contribution of this paper is to identify the effects of access regulation. We find
that prices are higher in countries where entrants make much greater use of bitstream
access, and lower in markets where they rely more heavily on direct access (local loop
unbundling, LLU). Despite this, we observe no significant effect on prices when entrants
upgrade their own networks, nor do we find any effect of inter-platform competition
between xDSL, cable and FTTx. These results might be interpreted as a consequence of
the application of a “ladder of investment” approach (LOI),
4
whereby in order to promote
sector competition regulators need initially to facilitate the access of new entrants to the
incumbent’s network so as to guarantee service-based competition, and subsequently, once
these entrants have acquired experience and reputation they need to provide incentives to
entrants to invest in their own infrastructure. In this way, the long-term benefits of facility-
based competition can be reconciled with short-term price reductions.
The rest of the paper is organised as follows. Section 2 reviews the economic literature, so
as to highlight the contributions of this paper, and it describes the European broadband
market. Section 3 outlines our estimation strategy. Section 4 describes the data set. Section
5 presents the empirical strategy and results and Section 6 discusses the main contributions.
Finally, Section 7 concludes.
2. Literature review and European broadband market
2.1 Broadband access: the literature
The initial empirical literature on broadband Internet access focussed on the determinants
of its penetration. For example, Distaso, Lupi and Manenti (2006) reported the impact of
inter-platform competition on broadband penetration in 14 European countries from 2000
to 2004. They found that while inter-platform competition had a positive effect on
penetration, intra-platform did not play an important role. Other studies, including Höffler
(2007), highlighted the inefficiencies created by the duplication of existing platforms.
More recent papers have analysed the impact of access policies on the investment decisions
of firms and on the diffusion of the service.
5
Grajek and Röller (2012) examine the effects
of access regulation on incentives for investment in 20 countries in the period 1997-2006.
They report that regulation discourages the investment of incumbents and individual
entrants, and suggest that the European regulatory framework fails to provide incentives
for facility-based competition. They also examine the regulators’ response to infrastructure
investments, concluding that access regulation is not affected by entrants’ investments, but
rather that the intensity of regulation reflects the incumbent’s investments. Bouckaert, Van
4
The “ladder of investment” regulatory model was first identified by Cave (2006). See Cambini and Jiang
(2009) for an extensive review of the literature on this topic and Bourreau et al. (2010) for a critical analysis of
this regulatory approach.
5
A detailed review of the theoretical literature on access charges in telecommunications can be found in
Laffont and Tirole (2000), Arsmtrong (2002), and Vogelsang (2006).
4
Dijk and Verboven (2010) investigate the influence of competition on broadband
penetration in a sample of 20 OECD countries. They consider three entry patterns adopted
by broadband operators: (1) inter-platform competition, where the incumbent xDSL
operators compete with infrastructure-based operators (e.g. cable and FTTH); (2) facility-
based intra-platform competition, in which entrants lease some unbundled local loop
elements, but have to invest in their own equipment and facilities (e.g. LLU and share
lines); and (3) service-based intra-platform competition, where entrants resell the
incumbent’s services (bitstream access/resale). According to these authors, only
infrastructure-based competition increases the penetration of the service, while the other
types have little effect.
Few papers have undertaken specific country studies. Pereira and Ribeiro (2010) examine
the competition between xDSL and cable operators in Portugal. They find that inter-
platform competition (mainly between xDSL and cable) increases the diffusion of Internet
thanks to both the higher coverage of broadband access and the existence of lower prices.
More recently, Nardotto, Valletti and Verboven (2012) have analysed the impact of
unbundling on broadband penetration in the UK during the period 2005-2010 using micro
level information. They find that LLU had little or no effect on broadband penetration,
although it increased the quality of the service in terms of average broadband speed. On
the other hand, they show that inter-platform competition from cable increased local
broadband penetration.
Many of the above results contrast with those reported by Gruber and Koutroumpis (2013)
who, when drawing on data from 167 countries between 2000 and 2010, find that inter-
platform competition is an impediment to broadband adoption. They conclude that
markets that focus specifically on one type of technology typically present a more rapid
adoption process than that experienced in multi-technology markets. They suggest that this
is because the latter type of competition requires the duplication of networks, which
increases costs and, ultimately, prices. In contrast, full retail unbundling increases the
diffusion of the service.
While there have been several studies of broadband penetration in recent years, the
determinants of the prices of the service have received much less attention.
6
Explanations
for this include the absence of consistent data, and the fact that broadband services are
highly varied and typically offered jointly with voice telephony and television. An exception
is provided by Wallsten and Riso (2010), who examine a data set comprising 30 OECD
countries between 2007 and 2009. They find that downstream speed had a positive effect
on prices, that broadband plans with bit caps were on average offered at lower prices than
unlimited plans with contracts, and that plans with contracts were typically less expensive
than those without. Their paper has several points in common with our study, but they do
not examine the impact of access regulation on retail prices.
Greenstein and McDevitt (2011) also analyse the economic value created by the diffusion
of broadband Internet provided via xDSL and cable in the United States. They do not have
direct information on prices, but create a price index that allows them to adjust the price to
the progressive improvement in service quality. Taking this into account, they show that
broadband prices in the US fell slightly during the period 2004-2009. This is a very
6
Galperin (2012) describes the evolution of broadband prices in Latin America.
5
different evolution to that of the prices of electronic products, including laptops and
printers, where the quality-adjusted price falls have been significant.
2.2 The European broadband market
In July 2011, the average penetration level of fixed broadband Internet access in the
European Union was 27.2%
7
. However, there were significant differences across member
States. For example, while the penetration levels in Netherlands, Denmark and France were
39.3, 38.5 and 33.9%, respectively, in Romania, Bulgaria and Poland they were 14.6, 15.6
and 16.4%, respectively (Figure 1).
8
Figure 1: Fixed broadband penetration and incumbent’s penetration in 2011 (%)
Source: European Commission (2011a).
Figure 2: Fixed broadband prices in 2011 (€ PPP)
Least expensive offer (all ISPs): Basket 4096 kbps-8192 kbps, 5GB or 20 hours/month
Source: European Commission (2011b).
7
The penetration of large screen mobile broadband subscriptions (using dedicated data cards or USB
modems) was 7.5 % in July 2011.
8
See European Commission Implementation Reports and European Commission (2011a and 2011b).
4 5 5
7
11
10 10
9
11
8
12 11 10
12
18
14 14
12
23
17
23
15
9
12
15 14
24
17
14,6 15,6 16,4
17.0
19,3
20,8
20,9 21,4 21,6
22,1
22,3
23,6
23,9 24.0
25,2
25,6 26,6
27,2
29,2
30.0
31,4 31,6 32,4
32,4 32,7
33,9
38,5
39,3
RO BG PL SK LV EL PT HU LT CZ IT IE SI ES CY AT EE EU FI MT LU BE UK SE DE FR DK NL
Incumbent's penetration Total fixed broadband penetration
11 12 12
15
17 17
19
20
22 22 23 23 23
25 26 28
29
20
31
33 33 34
20
28
20
31
54
11
14
9
19
19
DK LT LV RO AT SE MT DE EL NL HU BG PL SK FI EE LU BE IE IT CZ SI FR UK ES PT CY
Tari ff Li ne rental
6
In recent years, the prices of fixed broadband Internet access have fallen significantly,
which is quite remarkable if we consider that operators have significantly improved the
quality of their offers. They often allow consumers to migrate at no cost to other offers
providing higher download speeds. Moreover, there is a trend towards bundling
complementary services, such as broadband with fixed voice and TV, and more recently
with mobile telephony (voice and data). Such packages allow operators to attract new
consumers (the bundle being cheaper than the sum of the single services) and, at the same
time, to gain the loyalty of their subscribers.
This situation has not prevented significant price differences across European countries
(Figure 2).
9
These differences can be accounted for in part by technical and commercial
features, but they also reflect the level of competition in the market. Thus, while in 2011
the incumbent’s market shares in Finland, Cyprus and Luxemburg were 79, 73 and 72%,
respectively; in the UK, Bulgaria and Romania they were 29, 29 and 30%. Many member
States have four or five alternative operators, but other markets are much more
fragmented. For example, in Germany there are around 100 regional entrants, though the
incumbent retains a 46% market share. The objective of our paper is to analyse how
technical, commercial and regulatory factors affect prices and to understand how they
determine price differences across member States.
Broadband access can be provided via several technologies. In the period we study, the
most frequently employed system is xDSL followed by cable modem, but some operators
use FTTH or wireless technologies such as 3G, Wimax and satellite. Around 77% of the
fixed access lines in European countries use the xDSL technology, which explains why the
average speed is still quite low (around 10 Mbps) and why there is more intra-platform than
inter-platform competition.
Incumbent operators are usually vertically integrated (except in Sweden, the UK and Italy)
and use xDSL (although some use cable, for example Portugal). Most entrants use the
incumbent’s network to provide their services and have to pay a regulated access fee. Cable
operators have built their own infrastructure, but they also need to sign interconnection
agreements with incumbent operators because of their limited national coverage.
In the EU, National Regulatory Authorities (NRAs) set interconnection prices in order to
guarantee an adequate development of competition. There are two mandatory types of
access. Entrants can access the incumbent’s network directly (LLU) or indirectly
(bitstream). At the same time, the direct access can be of three types: complete unbundling
of the local loop, where entrants pay to use the incumbent’s access lines without any
restriction; shared LLU, where entrants use the high frequencies of the access lines to
provide broadband and incumbents use the low frequencies to provide voice telephony;
and, shared LLU without voice telephony (naked ADSL), which is similar to the previous
service but voice telephony is offered over the Internet (VoIP). The main advantages of
unbundling are, therefore, to allow entrants to offer a differentiated service and to develop
their own commercial policy.
9
In the EU, retail prices of broadband services are not regulated. However, national regulators periodically
assess whether there is a “margin squeeze” that reduces the profitability of entrants. This might occur, for
example, when the level of interconnection prices (wholesale access prices) make impossible to entrants from
matching the incumbent’s offers.
7
In the case of indirect access (bitstream), entrants can access the incumbent’s network at
two levels: at the ATM level (or Gig-ADSL), where there are several geographical
interconnections, and at the IP level (or ADSL-IP), which is more expensive because there
are fewer interconnection points.
Price regulation of all these access services is an essential instrument for promoting
competition and investment. Regulated access prices determine in which part of the
incumbent’s network the entrants will choose to invest and influence both retail prices and
service quality. In the EU, following the “ladder of investment” (LOI) regulatory model,
regulators set the prices of indirect and direct access (bitstream and LLU) in order to
provide incentives to entrants to invest progressively in their own equipment. In spite of
this, the empirical literature is still unclear about the effectiveness of this strategy (Hazlett
and Bazelon, 2005; Bourreau and Dogan, 2006; Waverman et al., 2007; Bacache et al.,
2011, and Grajek and Röller, 2012). As Bourreau et al. (2010) explain, the main problem of
the LOI is that once entrants obtain some profits with bitstream access, their incentives to
invest may not be so high, creating a “replacement effect”. Moreover, the simultaneous
presence of multiple access levels can hinder incentives to access higher rungs on the
investment ladder. Our paper contributes to the empirical literature on access regulation by
assessing how the degree of intensity of each type of entry in a country affects retail prices.
3. Estimation strategy
This section examines the prices of broadband Internet access in 15 European countries in
the period 2008-2011. After adjusting for the hedonic features of the operators’ plans, we
analyse the impact of several commercial strategies adopted by operators, including
bundling and market segmentation. Additionally, we assess how different entry patterns
(bitstream, LLU and own network) induced by access regulations affect retail prices.
We estimate a model for the prices of broadband residential plans (p
mit
), where m is the
offer, i is the country, and t is the time period. The explanatory variables that we use in the
estimation can be grouped into three blocks: (1) technical characteristics of the service; (2)
the operators’ commercial strategies and (3) measures of competition and regulation in the
country. We also control for the degree of penetration of the service in the country, as well
as fixed country and time effects. Specifically, we estimate the following model:
pmit = a
0
+ a
1
Speedmit +a
2
Speedmit
2
+a
3
Technologymit
Technical characteristics
+a
4
Bundlingmit +a
5
VoIPmit +a
6
Datacapmit + a
7
Volumecapmit
Commercial characteristics
+a
8
Incumbentmit. + a
9
PDispersionit +a
10
NOffersit+a
11
HHIPlatit + a
12
Bitstreamit + a
13
Directaccessit + a
14
Ownetworkit
Measures of competition and regulation
+a
15
Penetrationit + a
16
Regiont +a
17
Timet +e
mit
8
The prices charged by operators in each country may vary according to the quality of the
service and the access technology. In our model, Speed is the downstream speed specified in
the technical details of the operators’ plans. To account for a possible non-linear
relationship between price and speed, the equation also includes the square of the speed.
Technology is the access technology used to provide the service; this might be xDSL, cable,
or fibre (FTTH). We expect each technology to have a different effect on the price given
that they require different levels of investment and bandwidths, and because consumers
might have different “perceptions” of the quality of each technology.
The price equation also includes commercial practices that may be adopted by operators.
Bundling refers to a broadband service provided together with voice telephony and/or
television. The commercialisation of broadband with other services implies additional costs
for the operators that can justify a price increase. For example, to be able to offer television
services, operators must first reach agreements with TV channels and pay them a
compensation. The variable VoIP reflects the situation where the broadband service is
bundled together with voice telephony but provided over IP, which reduces the operators’
costs (naked xDSL).
10
Datacap is a dummy variable that shows if the plan includes a restriction in the user’s
downstream volume (consumers have to pay an extra charge when they exceed the
established cap), and Volumecap measures the volume of data that the user can download. A
priori, we expect the capped offers to be cheaper than those with unlimited downstream
capacity, and also for prices to increase with the volume of the download limit.
We examine a group of variables that reflect the competition in the national markets.
Incumbent is a dummy that identifies the different pricing policies of the incumbent and the
entrants. Incumbents may have market power thanks to certain advantages of reputation or
to the existence of consumer switching costs. They may also have cost advantages over
their rivals. Yet, European operators may be the incumbent in one country and the entrant
in another or others. Thus, the costs of incumbent operators are also related to their
presence in several countries and to their bargaining power with their equipment providers.
HHIPlat is the concentration Herfindahl-Hirschman Index (HHI) in terms of technology
shares. A higher HHIPlat would mean a high concentration of a particular technology in a
given country. As discussed in Section 2, the empirical literature is ambiguous with regard
to the effect of inter-platform competition on the diffusion of the service (see, for example,
Bouckaert, van Dijk and Verboven, 2010 and Gruber and Koutroumpis, 2013), and its
effect on prices is also unclear. Here, a factor that should be considered is that inter-
platform competition allows operators to differentiate their services, which might offset
price reductions due to higher competition between different platforms.
PDispersion measures the price dispersion of all the plans commercialised in the country,
and NOffers is the number of plans offered by the incumbent. Both variables are proxies of
the degree of price discrimination in the market. Incumbents can offer a large number of
plans to better respond to competitors that target specific groups of consumers, but they
can also segment the market if they have market power.
10
Our model does not consider if the consumer can subscribe separately to each service (“menu à la carte”)
or if the client has no option other than to contract the bundle (so-called Tying). The reason is that it is
difficult to gather this information and also the problem of multicollinearity between the variables Tying and
Bundling.
9
Finally, a principal objective of this study is to determine how different types of entry in the
country (bitstream, LLU and the development of entrants’ own networks) affect the
operators’ pricing strategies. Bitstream, Directaccess, and Ownnetwork are explanatory variables
that reflect the relative importance of these entry patterns. Notice that important
distinctions can be drawn between Ownnetwork and HHIPlat. Thus, while the former
identifies an entrant that bypasses the incumbent’s network (implying the duplication of
networks), the latter reflects the presence of different technologies in the country, though
not necessarily the duplication of networks, i.e., the market might be geographically
segmented by technologies
11
. The inclusion of these variables in the pricing equation is
useful to analyse the effects of access regulation, although there are other important aspects
that also determine the entry strategy of operators, including the investment that is needed
to deploy the equipment, the operators’ perceptions of consumer willingness to pay for
high-quality services, or the regulatory institutions in the country.
Unfortunately, our data set does not provide any information about the number of clients
that subscribed to each of the commercial plans. Yet, the variable Penetration offers details
of the number of subscribers in each country for five different speed ranges. In the
presence of economies of scale, we expect operators to set lower prices as they have a
larger penetration and more subscribers to their plans. However, this effect may be
moderated when the increase in penetration is achieved as a result of extending service
coverage to high cost/low density regions.
4. The data
We use a panel data of residential retail broadband offers in 15 European member States
for the period December 2008 to June 2011. The countries analysed represent more than
80% of the total broadband access lines in the EU. On average the data set contains
around 550 offers per year and an overall total of 2,204 observations (Table 1). The sample
includes the tariffs of operators whose accumulative subscribers represent over 90% of the
broadband market in each country. Most of our data are drawn from Quantum-Web Ltd.
Data for the countries’ penetration rates and socio-economic variables are provided by the
European Commission, Eurostat and UNESCO.
Information concerning the technical and commercial characteristics of each offer is
collected primarily from the operators’ web sites by Quantum-Web. It should be stressed
that the prices offered on these sites might differ in some cases from those offered via
other sales channels (e.g. operators’ retail shops). Likewise, operators may negotiate
discounts with consumers when the latter have received better offers from rivals or when
they want to attract new subscribers. In the case of the technical characteristics, note that
the download speed advertised might not be uniform across all households.
The price of the offer used as the dependent variable is the monthly payment in euros
advertised by the operator and adjusted by each country’s purchasing power parity (PPP).
Some offers include the monthly line rental fee, which is commonly associated with voice
telephony. This is typically the case for xDSL, unless the voice service is provided via
VoIP, but not in the cases of cable and FTTH technologies. To account for this situation,
11
An example of market segmentation by technology is Belgium where the broadband lines in Flanders are
usually cable, while in Wallonia there is a more intensive use of xDSL.
10
we consider the price of each offer as being the sum of the access fee (the line rental) and
the broadband service tariff
12
. Note that in some cases the offer establishes limits to the
downstream volume, which means that consumers have to pay an additional charge once
this limit has been exceeded. Our price does not include these additional payments,
although the variables Datacap and Volumecap control for this situation.
We use two hedonic variables to estimate the price. Speed is the quality of the service
advertised on the operators’ web pages and is measured in Mbps. The minimum speed in
our sample is 0.128 kbps and the maximum is 500 Mbps. However, a significant number of
plans offer a quality between 10 and 30 Mbps (Table 2).
Source: Quantum Web-Ltd
The variable Technology reflects the type of access used to provide the service; this can be
xDSL, cable or FTTH. Taking this into account, the variable takes a value of 1, 2 or 3,
respectively. It should be stressed that the download speed is related to the technology
used in providing the service. Thus, xDSL cannot offer more than 30 Mbps, with the sole
exception of VDSL which can reach 50 Mbps. By contrast, cable supports speeds of up to
100 Mbps (DOCSIS3.0) and FTTH can attain download speeds of 1 Gbps. Also, an
operator’s bundling strategy depends on its technology. xDSL is usually bundled with voice
telephony, but not with television because good quality cannot be guaranteed. Cable and
FTTH, on the other hand, are often provided together with television.
Broadband access can be bundled with other services and commercialised at a single price.
To identify the incremental price that this implies, the variable Bundle takes values of 1, 2, 3
or 4, depending on whether broadband access is offered individually or bundled with fixed
voice telephony, TV, or with both
13
.
12
Prices do not include any other costs, such as the price of routers or the installation fee.
13
In contrast with Wallsten and Riso (2010), we have no information about the number of channels in triple
play packages.
Table 1: Descriptive statistics. Period 2008-2011
Variable Observations Mean Standard
Deviation
Minimum
Values
Maximum
Values
Price (euros) 2204 35.8 14.8 7.1 138.5
Price Single broadband (euros) 909 30.3 12.3 7.3 82.5
Price Broadband and voice (euros) 699 35.9 12.9 7.1 107.7
Price Broadband and TV (euros) 116 39.7 12.9 15.1 72.2
PriceBroadband, voice and TV (euros) 480 45.2 16.8 13.8 138.5
Price for Metered offers (euros) 410 36.7 14.7 7.1 79.8
Volume cap (Gb) 410 64.0 135.9 0.4 1000.0
Speed (Mbps) 2204 23.8 32.6 0.1 500.0
HHI inter-platform 2204 63.6 17.2 38.0 100.0
Bitstream access Index 2204 4.2 8.3 0.0 48.27
Direct access Index 2204 32.1 44.7 0.2 171.7
Own network access Index 2204 52.1 76.5 0.0 405.9
11
Datacap is a dummy variable that takes a value of 1 when the offer includes a limit to the
downstream volume, and 0 otherwise. The variable Volumecap measures the limit to the
download capacity (rated in GB) in offers with a downstream volume. Consumers usually
pay an extra charge if they exceed the downstream limit, but as explained above we do not
consider these charges in our analysis
14
.
Competition and regulation are essential aspects for understanding the operators’ pricing
policy. Our dataset provides information about the number of lines per operator in each
country, classified according to the technology used and the type of access (bitstream,
direct access, and own network). We use this information to construct concentration
indexes that measure the entry patterns in each country. Bitstream access concentration
(Bitstream) is defined as the square of the entrants’ bitstream market share (GigADSL o IP-
ADSL) in the country divided by the square of the incumbent’s market share. Direct access
concentration (Directaccess) is the square of the entrants’ direct access market share divided
by the square of the incumbent’s market share. Own network concentration (Ownnetwork) is
the sum of the squares of the entrants’ market share of own lines divided by the square of
the incumbent’s market share. As such, these indexes show the relevance of alternative
entry patterns in relation to the incumbent’s market share. In most European countries,
broadband services are in the main provided by legacy communication infrastructure,
where the incumbent operator maintains significant market power. Taking this into
account, we analyse how the intensity of different entry patterns with respect to the
incumbent’s position affects prices, and we also examine the response of prices to each
entry pattern.
A further instrument for measuring the competition level is the concentration of
technological platforms. To capture this, HHIPlat is the Herfindahl-Hirschman Index for
each country, which is estimated by adding the sum of the squares of market shares by
technology (xDSL, cable, FTTx).
PDispersion and NOffers can be interpreted as proxies of the degree of price discrimination,
and are calculated with the information obtained by Quantum-Web from the operators’
web sites. PDispersion is the standard deviation of broadband prices, and is calculated as the
square root of the variance in prices in each country for each year. NOffers is the number of
offers commercialised by the incumbent in each country and in each year. We might
alternatively have used the total number of offers commercialised in the country, but have
opted not to do so as we are unsure that our data set contains information about all
existing entrants.
Penetration is defined as the number of broadband subscriptions per 100 inhabitants in a
country. For the level of penetration we use EU information for five downstream speed
ranges: (1) below 2 Mbps, (2) 2-9.99 Mbps, (3) 10-29.99 Mbps, (4) 30-99.99 Mbps, and (5)
above 100 Mbps (ultrafast speed). The last two ranges are usually provided by cable or
FTTx, although the VDSL can also support speeds up to 50 Mbps.
14
Metered broadband plans charge for an additional capacity (GB) and some offers are time metered. The
extra costs (the so-called overage charges) are paid per GB or per a discrete number of extra GB, and less
frequently the plan charges per day, hour or minute above the cap limit. In some cases, no overage charges
are made and the broadband plans experience a sharp reduction in download speed once the cap has been
exceeded (bandwidth throttling).
12
The pricing equation also includes the dummy variable Region to account for the
unobserved heterogeneity in each market. The EU-15 countries have been clustered into
three groups owing to very high multicollinearity between country dummies. We have
tested that there are no significant differences in prices within each group after controlling
for the rest of the variables. In this way we avoid problems of multicollinearity and the loss
of degrees of freedom in the regression. The Region variable takes a value of 1 for Spain,
Belgium, Greece, Luxembourg and Portugal; 2 for Austria, Denmark, Netherlands, Finland
and Ireland; and 3 for France, Germany, Italy, Sweden and United Kingdom. Finally, Time,
is a dummy fixed effect for each year.
For illustrative purposes, Table 2 shows some of the most relevant characteristics of the
broadband plans for each country in 2011. The table highlights the differences across
countries in terms of price and download speed. Direct observation of these statistics
suggests that average prices might be affected by the speeds, but also by other factors such
as bundling or volume caps. The econometric analysis conducted in the next section seeks
to identify the main factors determining the operator prices and sheds some light on the
differences across countries.
Source: Quantum-Web Ltd
5. Estimation and results
This section presents an econometric multivariate analysis of the factors influencing
broadband Internet access prices. We estimate the pricing equation using three estimation
procedures: ordinary least squares (OLS), two-stage least squares (2SLS-IV) and three-stage
least squares estimations (3SLS-IV)
15
. We believe that our OLS estimates may suffer from
an endogeneity problem. Indeed, economic theory suggests that in countries where prices
are low, penetration levels might be higher. In such a case, the OLS coefficients of
15
We have discarded the use of a fixed effects model because of the severe attrition or identification of our
broadband plans (offers change over time or are substituted by others) and the little variance in our
observations during the period considered (with information covering four years). We have also ruled out the
use of a random effects model because the unobserved heterogeneity (the firm or country unobserved
characteristics) will be correlated with the explanatory variables in the pricing equation.
Table 2: Residential Broadband Plans. Characteristics by Country in 2011
Observa-
tions
Number of
Operators
Average
Price
(euros)
Average
Download
Speed
Bundling
(% bundled
plans)
Metered
Offers (%)
Average
Volume Cap
(Gb)
Bitstream
market
share (%)
ULL market
share (%)
Own Network
market share
(%)
Austria 71 7 39.6 29.9 58% 20% 58.4 2% 13% 25%
Belgium 32 5 42.6 20.2 34% 56% 25.5 5% 4% 37%
Denmark 12 3 25.9 27.3 58% 42% 208.3 7% 10% 22%
Finland 26 4 26.4 27.5 0% 4% 0.1 3% 2% 63%
France 46 5 34.3 52.9 93% - - 8% 43% 7%
Germany 55 10 26.6 28.2 65% 2% 1.8 7% 36% 12%
Greece 33 5 40.3 18.6 61% - - 2% 56% 0%
Ireland 38 4 38.1 17.3 55% 63% 29.2 20% 5% 26%
Italy 26 6 29.1 10.6 42% 15% 0.1 14% 30% 4%
Luxembourg 19 3 36.3 18.6 32% 16% 1.4 0% 11% 0%
Netherlands 60 9 39.3 29.6 53% - - 2% 13% 40%
Portugal 33 5 53.1 69.4 88% 39% 18.2 2% 9% 40%
Spain 47 7 56.9 20.9 89% 2% 0.0 6% 24% 18%
Sweden 53 5 28.7 36.9 32% - - 4% 14% 39%
UK 39 6 30.7 23.4 67% 31% 7.0 11% 38% 21%
13
penetration are likely to be biased downwards, and so we would erroneously conclude that
penetration has a smaller effect on price than it actually has. Taking this into account, we
use instrumental variable techniques and we examine the possibility of using different
socio-economic variables as instruments for Penetration: GDPpc is the gross domestic
product per capita; Density is the number of inhabitants in the country divided by its area in
square kilometres; Education is the percentage of population with upper-secondary
education; and Unemployment is the unemployment rate in each country. Data for GDP,
Density and Unemployment have been obtained from Eurostat, and Education from the
UNESCO data base.
We expect GDPpc, Density and Education to have a positive effect on the adoption of
internet and Unemployment to have a negative effect. We expect GDPpc to affect Internet
penetration but not broadband prices. In addition, both Price and GDPpc are adjusted by
each country’s PPP so as to account for differences in the cost of living across EU member
States. Density should reduce operator’s costs for network deployment and, hence, should
have a positive effect on coverage. However, the effect of Density on price is unclear, since
while Density is defined at a national level, population density can vary greatly across
regions, which might affect the pricing decisions of incumbents and entrants. As for
Education, we expect the more highly educated to be more likely to contract Internet, but
the level of education should have no impact on the operators’ price setting decisions.
Likewise, it should be noted that the concentration indexes (HHIPlat, Bitstream, Directaccess
and Ownnetwork) might also be affected by an endogeneity problem since the entrants’ entry
decisions could be determined simultaneously with prices. Yet, a high index value might
also reflect the presence of efficient entrants that are able to provide the service at a low
price, or who are perceived by consumers as being better. To account for this situation, our
model includes country fixed effects to represent the unobserved characteristics that
influence the efficiency of operators and, eventually, the level of retail prices. Examples of
these unobserved effects include investments, administrative constraints, and state aid
plans that are specific to each country.
Table 3 reports the OLS and 2SLS/3SLS-IV estimates of the pricing equation. The first
three columns present the OLS results. Specification 1 considers the technical
characteristics of the offers and the commercial strategies of the operators. Specification 2
includes the competition and regulatory variables and specification 3 adds Penetration.
Finally, specifications 4 and 5 show the results of the 2SLS and 3SLS estimations using as
instruments the socio-economic variables GDPpc, Density and Education. Since the
dependent variable price is included in logs, Penetration is interpreted as a semi-elasticity. All
specifications include the Time and Regional Effect dummy variables.
14
The estimates of the pricing equation are robust to the alternative specifications. Most of
the coefficients in the regressions are significant and their signs are in line with our
predictions.
16
As expected, Speed increases broadband prices. Specifically, a 10 Mbps
increase in speed raises broadband prices by an average of around 5%. However, the
negative sign of Speed
2
reveals that price exhibits diminishing returns with respect to speed.
Additionally, after controlling for other characteristics of the plans, xDSL appears to be
more expensive than cable and FTTH. Fibre and cable technologies can provide higher
speeds and better quality than xDSL, but it seems that this is not sufficient to enable
entrants to charge higher prices per Mbps than are charged for xDSL. Such a situation can
constitute an obstacle to the authorities’ objective of promoting investment in Next
Generation Access Networks (NGAs). This finding might be justified by the fact that
xDSL is often the only available technology in many regions and locations and as a
consequence incumbents fix uniformly higher prices for low speed offers per Mbps than
those set by cable and fibre operators, who tend to be present only in densely populated
areas where there are several competitors.
As for the operators’ commercial strategies, unlimited plans (flat rates) charge higher prices
than metered plans. In the case of metered offers, when the volume cap is exhausted the
average increase in the price for an extra GB is slightly lower than 0.05%. On the other
hand, bundles of services are more expensive than stand-alone broadband plans. Offers
combining broadband access and TV are more expensive than those that bundle
broadband access and voice telephony. Specifically, broadband packages with voice or TV
are 15 and 18% more expensive, respectively. When the offer includes both voice
16
We test for multicollinearity using the variance inflator factor (VIF). Its mean equals 1.9 and the two highest
values are 4.9 for Speed and 3.4 for Speed
2
; the rest of the VIF values are below 3.
Table 3: Estimation Results: All Broadband plans. OLS; 2SLS/3SLS-IV
Dependent variable Specification 1 Specification 2 Specification 3 Specification 4 Specification 5
Log Price (price) OLS OLS OLS 2SLS-IV 3SLS-IV
Independent variables Coefficient Coefficient Coefficient Coefficient Coefficient
Penetration -0.0057* (0.00312) -0.017** (0.00723) -0.019** (0.00723)
Speed 0.0065*** (0.00146) 0.0069*** (0.0014) 0.006*** (0.00159) 0.0042*** (0.00185) 0.0047*** (0.00107)
Speed² -1.1e-05** (0.000004 -1.1e-05** (0.000004 -9.2e-06* (0.000004) -5.4e-06* (0.000005) -1.5e-05* (0.000003)
Technology dummy (reference xDSL)
Cable -0.092* (0.05150) -0.11** (0.04710) -0.11** (0.04577) -0.15*** (0.03449) -0.10** (0.02133)
FTTx -0.11** (0.0582) -0.11** (0.05026) -0.12** (0.05135) -0.14** (0.04405) -0.11** (0.02515)
Metered offer (Data cap) -0.098 (0.05860) -0.17*** (0.05739) -0.18*** (0.05816) -0.18*** (0.05773) -0.16*** (0.01878)
Volume cap 0.00043 (0.00030) 0.00049* (0.00023) 0.00049** (0.00022) 0.00048** (0.00021) 0.00042** (0.0001)
Bundle (refrence stand alone broadband)
Internet and voice 0.18*** (0.02848) 0.17*** (0.02719) 0.17*** (0.02591) 0.17*** (0.02372) 0.15*** (0.01482)
Internet and tv 0.19*** (0.04024) 0.20*** (0.03544) 0.20*** (0.03318) 0.21*** (0.03008) 0.18*** (0.02804)
Internet, voice and tv 0.36*** (0.05198) 0.34*** (0.04016) 0.34*** (0.03795) 0.34*** (0.03351) 0.31*** (0.01743)
VoIP -0.073 (0.04870) -0.077 (0.04668) -0.072 (0.04296) -0.063 (0.03831) -0.078*** (0.0232)
Incumbent 0.15*** (0.03165) 0.14*** (0.03282) 0.14*** (0.03265) 0.12*** (0.03099) 0.15*** (0.01836)
HHIPlat 0.0014 (0.00110) 0.00089 (0.00116) -0.00016 (0.00080) -0.00035 (0.00113)
Bitstream 1.1*** (0.25199) 1*** (0.23629) 0.96*** (0.21353) 0.99*** (0.11317)
Direct Access -0.16*** (0.05061) -0.14** (0.04868) -0.092* (0.04770) -0.073** (0.03442)
Ownnetwork 0.004 (0.01470) -0.00062 (0.01290) -0.0095 (0.01254) -0.0099 (0.01112)
PDispersion 0.022*** (0.00710) 0.02*** (0.00625) 0.017*** (0.00530) 0.02*** (0.00327)
Noffers Incumbent 0.0023 (0.00222) 0.0043* (0.00225) 0.0081** (0.00409) 0.0026 (0.00289)
R
2
0.48 0.52 0.53 0.51 0.49
Number of observations (N) 2204 2204 2204 2204 2204
Note: Robus t standard errors are i n parenthes i s . Si gni fi cance at * 10%, ** 5%, *** 1% l evel s .
15
telephony and television, prices rise by about 30%. By contrast, offers that include
broadband and voice can be up to 7% cheaper if the voice service is supplied over IP
(VoIP).
In the case of the competition variables, incumbents’ plans are around 15% more
expensive than entrants’ offers, which might reflect the formers’ dominant position in the
market and “unobserved” quality differences as perceived by consumers. PDispersion and
Noffers exhibit a positive effect on prices, which suggests that when firms are better able to
screen consumers they set higher prices, and that a large number of plans of varying
characteristics and prices might generate a certain confusion among consumers, thus,
allowing firms to set higher prices (Hoernig, 2001). We also find that technological
concentration, measured with the variable HHIPlat, does not have a significant effect on
prices.
Our estimates show the entry pattern to be an important determinant of broadband prices
in a country. While the intensity of Bitstream entry increases prices, Direct Access (LLU)
reduces them, and Ownetwork does not entail a significant effect. On the other hand, the
coefficient associated with Bitstream is higher than that associated with Direct Access,
indicating that with an equivalent variation in the index there will be a greater price reaction
with Bitstream. One explanation is that LLU allows operators to differentiate their products
and to develop their own commercial strategies, which may avoid large price reductions for
equivalent levels of entry.
As expected, the OLS estimate shows that the Penetration variable has a negative effect on
prices. However, OLS may produce biased estimators if price and penetration are
determined simultaneously. To account for this, we applied the instrumental variable
approach (2SLS) using as our instrument a combination of the previously defined socio-
economic variables GDPpc, Density, Education and Unemployment.
Table 4 presents the result of the Hausman test that confirms that Penetration is an
endogenous variable. It also shows the results of the validity tests for different
combinations of instruments. The instruments in the second and third specification pass
the Hansen’s J test for over-identifying restrictions and are also shown to be suitable for
explaining Penetration. However, the validity test results for GDPpc and Unemployment (Test
Result 1) are not satisfactory. Moreover, Hansen’s J test value for the third specification is
lower than the second value. Therefore we use GDPpc, Density and Education as our
instruments.
Finally, we apply the instrument suitability tests (the F-statistic in the first stage regression
of the variable Penetration) to verify that the instruments are strong. Additionally, the
computed standard errors are robust to any bias from heteroskedasticity and they are also
clustered according to observations from the same country.
Table 4: Endogeneity test for Penetration
Hausman endogeneity test. Ho: Penetration exogenous p-value= 0.0000
Instruments for Penetration GDPpc, Unemployment
GDPpc, Education,
Density
GDPpc, Education,
Density, Unemployment
Test Test Result 1 Test Result 2 Test Result 3
Hansen J test. Ho: instruments exogenous p-value=0.1162 p-value=0.1691 p-value=0.1140
Validity of Instruments Ho: weak instruments p-value=0.1489 p-value=0.0001 p-value=0.0001
16
Usually, in simultaneous equation models (SEM), the variables determined simultaneously
(in this case, Price and Penetration) are correlated with the error term in the regression,
causing the OLS coefficients to be biased and inconsistent. To overcome this problem, we
have also estimated the pricing equation using 3SLS, which is a more efficient method
17
.
The results of the 2SLS and 3SLS estimations are shown in columns 4 and 5 in Table 3. We
find that the technical, commercial, and competition coefficients are robust to the use of
these estimation techniques. In the case of Penetration the sign of the coefficient is
maintained, but its value is higher than that obtained with the OLS regression. In the case
of the 3SLS-IV regression, the coefficient ??
P-3SLS
= -0.019 implies that a one-percentage-
point increase in the penetration level is followed by a 1.9% fall in price. The negative
estimate of Penetration instrumented through the socio-economic variables more than triples
the OLS coefficient, showing that the OLS Penetration coefficient was downward biased.
We also considered the effect of using the lagged Penetration variable as our instrument,
where LagPenetration and Lag2Penetration are defined as the variable lagged one and two
periods, respectively. Table 5 shows the Hausman endogeneity test for the lags of
Penetration. Both variables are endogenous but their Hausman chi-square test values are
lower than that for Penetration. The latter can be accounted for by the fact that prices and
the subscription decision are determined in the same period and, as such, the simultaneity
problem should be reduced for lagged periods of penetration. We ran the same OLS
regression as specification 3 in Table 3 and found that the coefficient is ??
LagPenetration-OLS
= -
0.007 when we consider one lag and ??
Lag2Penetration-OLS
= -0.0076 when we consider two lags of
Penetration, and in the two cases the coefficients are significant at the 5% level. These results
confirm that the simultaneity bias is downwards.
6. Discussion
Below we discuss in greater detail two key features that affect the operators’ commercial
policies: their bundling strategies and entry patterns.
Bundling strategies - A commercial strategy widely adopted by telecom operators in the
EU is that of bundling several services together in the same offer. Our analysis shows how
each additional service in such a bundle increases the price of the broadband plan.
However, standalone broadband and bundled offers can be considered as different
products and operators price them differently. To account for this possibility, we present
17
The 3SLS estimation of the model takes into account the structural equations for both Price and Penetration
and uses the cross-equation correlation of errors to provide more efficient estimators than the 2SLS
estimation. The procedure involves an additional step to the 2SLS procedure: namely, it uses the residuals
from both the Price and Penetration structural equations to construct a covariance matrix of errors that
provides a more efficient estimation.
Table 5: Endogeneity test for Penetration and Lags of Penetration
Hausman endogeneity test. Ho: Penetration exogenous Chi2(1)= 38.345 (p-value = 0.0000)
Hausman endogeneity test. Ho: LagPenetration exogenous Chi2(1)= 31.801 (p-value = 0.0000)
Hausman endogeneity test. Ho: Lag2Penetration exogenous Chi2(1)= 7.725 (p-value = 0.0054)
17
separate estimations for standalone broadband and for the bundle of broadband and voice
services
18
. Table 6 shows that these new estimations are very similar to those found when
we consider all offers. In the 2SLS and 3SLS estimations, Penetration is still instrumented by
the socio-economic variables GDPpc, Density, and Education, but its coefficient is no longer
significant
19
. By contrast, HHIPlat is now significant, indicating that a higher concentration
of one technological platform (less inter-platform competition) raises prices. In the case of
bundles, this result might be further reinforced because cable and FTTH are more likely to
offer TV than is the case with ADSL. As a result, technological concentration generates
higher prices for bundled offers.
18
In the case of bundle offers, we are unable to identify the price of each particular service, but we can
observe how the price of an offer increases as other services are integrated. Wallsten and Riso (2008) also
adopt this approach when analyzing bundling.
19
The penetration information we use is based on the whole sample since it is not possible to distinguish
between penetrations rates that depend on bundled plans, on the one hand, and those that depend on
unbundled plans, on the other.
Table 6: Estimation Results: Standalone Broadband and Bundles. OLS; 2SLS/3SLS-IV
Dependent variable
Log Price (price) OLS 2SLS-IV 3SLS-IV OLS 2SLS-IV 3SLS-IV
Independent variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient
Penetration -0.0019 0.0036 0.0038 -0.0038 -0.0041 -0.0057
(0.00331) (0.00824) (0.00919) (0.00229) (0.00831) (0.00899)
Speed 0.0088*** 0.0096*** 0.01*** 0.0071*** 0.0070*** 0.0075***
(0.00185) (0.00160) (0.00144) (0.00185) (0.00239) (0.00215)
Speed² -1.5e-05* -1.6e-05* -1.9e-05*** -1.7e-05* -1.6e-05* -2.3e-05***
(0.000005) (0.000005) (0.000003) (0.000008) (0.000008) (0.000008)
Technology dummy (reference xDSL)
Cable -0.18*** -0.17*** -0.12*** -0.11** -0.11** -0.08***
(0.05689) (0.05295) (0.02962) (0.05119) (0.04786) (0.02969)
FTTx -0.17** -0.15** -0.09** -0.086 -0.087* -0.057**
(0.07424) (0.06814) (0.04334) (0.06263) (0.05163) (0.03739)
Metered offer (Data cap) -0.15* -0.14** -0.1*** -0.2** -0.2*** -0.18***
(0.07203) (0.07050) (0.03117) (0.07460) (0.07306) (0.02791)
Volume cap 0.0003 0.0003 0.00022 0.0006** 0.0006*** 0.00055***
(0.00030) (0.00029) (0.00014) (0.00023) (0.00021) (0.00015)
VoIP -0.038 -0.038 -0.043
(0.03779) (0.03572) (0.02913)
Incumbent 0.11** 0.12** 0.14*** 0.18*** 0.18*** 0.2***
(0.04379) (0.04662) (0.02936) (0.04249) (0.03574) (0.02773)
HHIPlat 0.0016* 0.0021*** 0.0024** 0.0042* 0.0042* 0.004**
(0.00077) (0.00066) (0.00113) (0.00217) (0.00218) (0.00131)
Bitstream 0.89*** 0.93*** 0.96*** 0.86*** 0.86*** 0.86***
(0.27161) (0.22084) (0.18773) (0.21209) (0.20025) (0.14252)
Direct Access -0.1 -0.12** -0.11** -0.17*** -0.17*** -0.16***
(0.06613) (0.05136) (0.04758) (0.03608) (0.03444) (0.04247)
Ownnetwork -0.0058 0.00057 0.0036 -0.019 0.02 0.022
(0.01037) (0.01428) (0.01491) (0.08236) (0.08320) (0.05283)
PDispersion 0.023*** 0.026*** 0.029*** 0.022*** 0.022*** 0.022***
(0.00649) (0.00711) (0.00568) (0.00693) (0.00734) (0.00458)
Noffers Incumbent -0.00044 -0.0028 -0.0094* 0.0075* 0.0075* 0.0052*
(0.00383) (0.00372) (0.00487) (0.00412) (0.00415) (0.00307)
Regional Effect (reference Region 1)
Region 2 -0.21*** -0.22*** -0.21*** -0.24*** -0.24*** -0.24***
Region 3 -0.41*** -0.41*** -0.39*** -0.42*** -0.42*** -0.42***
Time dummy (reference t=2008)
t=2009 0.071 0.069 0.064** 0.061 0.061 0.059**
t=2010 0.081 0.079* 0.072** 0.12** 0.12** 0.12***
t=2011 0.11** 0.11** 0.097*** 0.15** 0.15** 0.15***
Cosntant 3.1*** 2.9*** 2.9*** 3*** 3.1*** 3.1***
R
2
0.42 0.42 0.40 0.58 0.58 0.57
Number of observations (N) 909 909 909 699 699 699
Note: Robust standard errors are i n parenthesi s. Si gni fi cance at * 10%, ** 5%, *** 1% l evel .
Stand-alone Broadband Bundle: Broadband + Fixed Voice Telephony
18
We are also interested in identifying the factors that serve as incentives for operators to
commercialise bundled services, since significant differences can be observed across
European countries (Table 2). In order to address this question we run logit and probit
discrete choice models that consider 1,336 observations collected on a yearly basis between
2008 and 2011. This regression focuses only on xDSL offers since bundling decisions differ
substantially for cable and FTTH. For this sample, 42% of the offers are broadband
standalone offers, 37% are bundles of broadband and voice, and 21% include TV,
broadband and voice.
The results of the estimation for Bundle are presented in Table 7. As expected, Speed
increases the probability of broadband being bundled with other services. Clearly, high
download speeds are necessary in order to offer TV. However, no significant distinction is
observed between the bundling strategies of incumbents and entrants. Market and
technological concentration (HHIOper and HHIPlat) are negatively associated with the
probability of bundling. This suggests that bundling is used by operators to attract
consumers in competitive markets, whereas in more concentrated markets operators prefer
to sell additional services separately. We also consider the effects of price dispersion and
find that it increases the probability of bundling. Therefore, in markets with a wide-range
of prices, operators might commercialise bundles of services to differentiate their plans.
Finally, no correlation is found between bitstream and direct access and bundling, although
bundling is found to be less likely in markets where more entrants bypass the incumbent’s
network (own network).
Table 7: Logit/Probit for Bundling
Bundling Logit Probit
Independent variables Coefficient Coefficient
Speed 0.033*** (0.00827) 0.02*** (0.004516)
Incumbent 0.38 (0.2556) 0.23 (0.15672)
HHIOperator -0.093** (0.03781) -0.057** (0.02286)
HHIPlat -0.035** (0.01614) -0.022** (0.00998)
PDispersion 0.062* (0.03288) 0.038* (0.01957)
Bitstream -0.9 (0.7889) -0.56 (0.4786)
Direct Access 0.1 (0.4771) 0.075 (0.2862)
Own network -2.8*** (0.8134) -1.7*** (0.5001)
GDPpc 0.003 (0.0053) 0.0022 (0.003317)
Education -0.017 (0.01455) -0.011 (0.00888)
Density 0.00072 (0.00059) 0.00045 (0.00037)
Regional Effect (ref: Region 1)
Region 2 -0.3 (0.3644) -0.19 (0.2239)
Region 3 0.22 (0.4141) 0.14 (0.2555)
Constant 5.8* (2.3438) 3.6** (1.4299)
Pseudo R
2
0.11 0.11
Number of observations (N) 1336 1336
Dependent variable
Note: Robust standard errors are i n parenthesi s.
Si gni fi cance at * 10%, ** 5%, *** 1% l evel .
19
Entry pattern - One of the main results of our analysis is that broadband prices are
negatively related with LLU entry intensity, and positively related with bitstream entry
intensity. This result implies that during the period analysed intra-platform facility-based-
competition was more effective in reducing prices than was intra-platform service-based-
competition, once we control for other characteristics of the offer. At this juncture, it
should be stressed that the intensity of each country’s entry pattern is influenced by the
regulation of access charges. In fact, European regulators usually follow the LOI approach,
which proposes setting higher prices for bitstream so as to induce entrants to use direct
access (Cave, 2006; Höffler, 2007, Bourreau, Dögan and Manant, 2010). As shown in
Bacache, Bourreau and Gaudin (2011), this regulatory strategy may have facilitated the
migration from bitstream access lines to LLU, although their study does not find any ladder
effect between LLU and entrants deploying their own network.
The regulation of access charges in the EU has been designed to provide incentives for
investment by entrants; however, in this paper we have also shown that it affects retail
prices. Prices are higher in countries where entrants make greater use of bitstream entry
and lower in countries where they make greater use of direct access. In fact, while the
establishment of high access charges for bitstream might act as an incentive to incumbent
operators to set higher uniform prices for their Internet plans, the establishment of low
access charges for direct access might not result in an equivalent response by operators,
since this type of entry requires major investment and because it allows operators to
differentiate their product (Nardotto, Valletti and Verboven, 2012). This situation is
reflected in our estimates of the Bitstream and Directaccess concentration indexes, since the
value of the coefficient is much greater for the former than for the latter. For instance, in
specification 5, ??
Bitstream-3SLS
=0.99 and ??
Directaccess-3SLS
=-0.073, which reflects the greater increase
in prices due to bitstream concentration than the relative decrease due to LLU
concentration.
Our results also suggest that inter-platform competition does not affect significantly retail
prices. Entrants that invest in cable and FTTH networks are able to commercialise high-
quality services, but this does not necessarily mean a reduction in their retail prices.
Moreover, although cable and FTTH plans involve lower prices per Mbps than those
charged by xDSL plans, the former typically offer more downstream speed and additional
services such as TV, which increase prices. A further important aspect that should be taken
into account when interpreting our results is that in our model competition is introduced at
the national level, but in some countries technologies are geographically segmented and so
there is little competition between them.
6. Conclusions
This paper has analysed the determinants of the prices of broadband Internet access in 15
countries of the EU between 2008 and 2011. Our econometric model has focused on three
types of variable: (1) the technical characteristics of the offers; (2) the operators’
commercial strategies; and (3) the regulation and competition in the country. Besides, we
have controlled for the potential endogeneity of broadband penetration by using the
instrumental variable approach (2SLS/3SLS) and employing as instruments a group of
socio-economic variables that determine the demand for broadband services in each
country.
20
Our analysis has revealed that downstream speed is a significant driver of price: a 10 Mbps
increase in the download speed causes prices to rise by an average of around 5%.
Additionally, we have found that the price per Mbps paid for the service via cable or fibre
technologies is lower than that provided by xDSL, although these technologies usually
provide higher download speeds. On the other hand, the xDSL service usually involves
more intra-platform competition and has a wider coverage than other technologies. In this
context, an important question that emerges is whether consumer willingness to pay for
fibre and cable plans is sufficiently high to encourage operators to invest in NGAs.
The operators’ marketing strategies also play an important role in determining prices. When
the broadband service is bundled with voice telephony, the price increases by around 15%
and when it is bundled with both voice telephony and television it increases by around
30%. By contrast, consumers that contract the voice service through VoIP obtain
significant price reductions. We have also shown that operators are less likely to bundle
their offers when the market is less competitive or when there is less inter-platform
competition.
This paper has also contributed to the literature that analyzes how regulatory policy can
influence the development of the broadband market. Here, we have shown that broadband
prices are higher in countries where entrants make greater use of bitstream access and
lower in those countries making greater use of LLU. We have found no evidence that inter-
platform competition and stand-alone entry (the last rung on the “ladder of investment”
approach) reduce prices. This means that when entrants bypass the incumbents’ networks,
retail prices are not reduced. This might be because this type of entry does not generate
sufficient competition, and because the operators offer high-quality products that are more
expensive. All in all, our results can be interpreted as an assessment of the “ladder of
investment” approach, and they show the benefits of facilitating migration from bitstream
to LLU entry.
One limitation of our study is that we have not considered mobile broadband plans offered
via smartphones or dongles. Mobile broadband demand is boosting and future research
needs to consider the impact of the firms’ commercial strategies for mobile broadband
offers. For example, a rising number of operators are currently offering bundled mobile
and fixed broadband packages, which minimises the consumer churn rate (Prince and
Greenstein, 2011).
21
7. References
Andrés, L., D. Cuberes, M. Diouf and T. Serebrisky (2010), “The diffusion of the Internet: A cross-
country analysis, Telecommunications Policy, 34, 323-340.
Armstrong, M. (2002), “The theory of access pricing and interconnection”, in M. Cave, S.
Majumdar, and I. Vogelsang (eds.), Handbook of Telecommunications Economics, Vol. 1. Amsterdam:
Elsevier.
Bouckaert, J., T. Van Dijk and F. Verboven (2010), “Access regulation, competition, and
broadband penetration: An international study”, Telecommunications Policy, 34, 661-671.
Bacache, M., M. Bourreau and G. Gaudin (2011), “Dynamic Entry and Investment in New
Infrastructures: Empirical Evidence from the Telecoms Industry”, Working Papers in Economics and
Social Sciences, 11-01.
Bohlin, E. and O. Teppayayon (2010), “Functional separation in Swedish broadband market: Next
step of improving competition”, Telecommunications Policy, 34, 375-383.
Bourreau, M. and P. Dögan (2006), “Build or Buy Strategies in the Local Loop”, American Economic
Review, Papers and Proceedings, 96 (2), 72-76
Bourreau, M., P. Dögan and M. Manant (2010), “A critical review of the ladder of investment
approach”, Telecommunications Policy, 34 (11), 683-696.
Cambini, C. and Y. Jiang (2009), “Broadband Investment and Regulation: A literature Review,”
Telecommunications Policy, 33 (10-11), 559-574.
Cava, I. and A. Alabau (2006), “Broadband policy assessment: a cross national empirical analysis”,
Telecommunications Policy, 30, 445-463.
Cave, M. (2006), “Encouraging infrastructure competition via the ladder of investment”,
Telecommunications Policy, 30, 223–237.
Czernich, N., O. Falck, T. Kretschmer and L. Woessmann (2011), “Broadband Infrasctructure and
Economic Growth”, The Economic Journal, 121 (May), 505-532.
Cincera, M., L. Dewulf and A. Estache (2012), “On the (In)Effectiveness of Policies to Promote
Broadband Diffusion in Europe (2003-2010): An Economic Assessment, Ecore Discussion Paper.
Deligiorgi C., C.Michalakelis A., Vavoulas and D. Varoutas (2008), “Nonparametric estimation of a
hedonic price index for ADSL connections in the European market using the Akaike Information
Criterion”, Telecommunication System (2007) 36: pp173-179.
Distaso, W., P. Lupi and G. Manenti (2006), “Platform Competition and Broadband Uptake:
Theory and Empirical Evidence”, Information Economics and Policy, 18 (1), 87-106.
European Commission (2009).“15th Progress Report on the Single European
Electronic Communications Market – 2009“.
European Commission (2011a), “Broadband access in the EU: situation at 1 July 2011, Information
Society and Media Directorate-General 2009, COCOM09-29 Final.
22
European Commission (2011b), “Broadband Internet Access Costs (BIAC)”. Final Report,
Information Society and Media Directorate-General.Van Dijk – Management Consulting.
Galperin, H. (2012), “Precios y calidad de la banda ancha en América Latina: Benchmarking y
tendencias”, Documentos de Trabajo, Universidad de San Andrés.
Grajek, M., and L.-H. Röller (2012), “Regulation and investment in network industries: Evidence
from European telecoms”, Journal of Law and Economics, 55 (1): 189-216.
Greenstein, S. and R. McDevitt (2011), “The Broadband Bonus: Estimating broadband Internet’s
economic value, Telecommunications Policy, 35 (7) , 617-632.
Greenstein, S. and R. McDevitt (2011), “Evidence of a modest price decline in US broadband
services”, Information, Economics and Policy, 23 (2), 200-211.
Greenstein, S. and J. Prince (2011), “Does service Bundling reduce churn”, mimeo.
Gruber, H. and P. Koutroumpis (2013) Competition enhancing regulation and diffusion of
innovation: the case of broadband networks, Journal of Regulatory Economics, 43: 168-195.
Guerrero de Lizardi, Carlos (2006): “Una aproximación al sesgo de medición del precio de las
computadoras personales en México”, Economía Mexicana.
Hazlett, T. and C. Bazelon (2005), “Regulated Unbundling of Telecommunications Networks: A
Spetting Stone to Facilities-Based Competition”, mimeo.
Hoernig, S. (2001), “Collusion and confusion?”, mimeo.
Höffler, F. (2007), “Cost and benefits from infrastructure competition. Estimating welfare effects
from broadband access competition”, Telecommunications Policy, 31(6-7), 401-418.
Houpis, G., J. Lucena and J. Santamaría (2011) “Geographic Segmentation of Broadband Markets:
Appropiate Differentiation or Risk to a Single EU Market?”, Communications & Strategies, 82,
105-.
Laffont, J. and J. Tirole (2000), “Competition in Telecommunications”, Cambridge, Mass, The MIT Press.
Lee, S., M. Marcu and S. Lee (2011), “An Empirical analysis of fixed and mobile broadband
diffusion”, Information Economics and Policy, 23, 227-233.
Miravete, E. and L. Röller (2004), “Estimating markups under nonlinear pricing competition”,
Journal of the European Economic Association, 2, 526-536.
Nardotto, M., T. Valletti. and F. Verboven. (2012), “Unbundling the incumbent: Evidence from
UK broadband”, mimeo
Pereira, P. and T. and Ribeiro (2010), “The impact on broadband access to the Internet of the dual
ownership of telephone and cable networks”, International Journal of Industrial Organization,29, 283–
293.
Renda, Andrea, “Competition-regulation interface in telecommunications: What’s left of the
essential facility doctrine”. Telecommunications Policy, 34, 23-35.
Rosen, S (1974), “Hedonic Prices and Implicit Markets: Product Differentiation in Pure
Competition”, Journal of Political Economy, 34-55.
23
RTR (Austrian Telecoms Regulator).“Further Information about fixed-mobile Broadband
substitution in Austria (2010)”. RTR Working Paper.
Savage, S. and D. Waldman (2004), “United States Demand for Internet Access”. Review of Network
Economics, Vol.3, Issue 3 – September.
Schwarz, A. (2007), “Wholesale market definition in telecommunications: The issue of wholesale
broadband access”. Telecommunications Policy, 31, 251-264.
Vareda, J. and S. Hoernig (2007), The Race for Telecoms Infrastructures Investment with Bypass:
Can Access Regulation Achive The first best? CEPR Discussion Papers 6203.
Vogelsang, I. (2003), “Price Regulation of access to telecommunications networks”, Journal of
Economic Literature, 41 (3).
Wallsten, S. and J.L. Riso (2008), ”Residential and Business Broadband Prices: Part 1: An Empirical
Analysis of Metering and Other Price Determinants”, mimeo.
Waverman, L., M. Leschi, B. Reillier and K. Dasgupta (2007), “Access regulation and Infrastructure
Investment in the Telecommunications Sector: An empirical Investigation”, mimeo.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 26
Research Institute of Applied Economics Working Paper 2013/09 pag. 26
26
Llista Document de Treball
List Working Paper
WP 2013/09 “Broadband prices in the European Union: competition and commercial strategies” Calzada,
J. and Martínez, F.
WP 2013/08 “Double Penalty in Returns to Education: Informality and Educational Mismatch in the
Colombian Labour market” Herrera-Idárraga, P.; López-Bazo, E. and Motellón, E.
WP 2013/07 “Firm exports, innovation, … and regions” López-Bazo, E. and Motellón, E.
WP 2013/06 “Fiscal sustainability and fiscal shocks in a dollarized and oil-exporting country: Ecuador”
Marí Del Cristo, L.M. and Gómez-Puig, M.
WP 2013/05 “Informality and Overeducation in the Labor Market of a Developing Country” Herrera-
Idárraga, P.; López-Bazo, E. and Motellón, E.
WP 2013/04 “GLS based unit root tests for bounded processes” Carrion-i-Silvestre, J.Ll. and Gadea,
M.D.
WP 2013/03 “Determinants of Broadband Access: Is Platform Competition always the Key Variable to
Success?” Fageda, X.; Rubio, R. and Termes, M.
WP 2013/02 “Beyond Value-at-Risk: GlueVaR Distortion Risk Measures” Belles-Sampera, J.; Guillén, M.
and Santolino, M.
WP 2013/01 “How systemic is Spain for Europe?” Claeys, P. and Vašícek, B.
WP 2012/24 “Changes in Wage Structure in Mexico Going Beyond the Mean: An Analysis of Differences
in Distribution, 1987-2008” Tello, C.; Ramos, R, and Artís, M.
WP 2012/23 “Decomposing the Rural-Urban Differential in Student Achievement in Colombia Using
PISA Microdata” Ramos, R.; Duque, J.C. and Nieto, S.
WP 2012/22 “Recovery Risk and Labor Costs in Public-Private-Partnerships: Contractual Choice in the
U.S. Water industry” Albalate, D, Bel, G. and Geddes, R.
WP 2012/21 “Beyond pure public and pure private management models: Mixed firms in the European
Airport Industry” Albalate, D.; Bel, G. and Fageda, X.
WP 2012/20 “The determinants of contractual choice for private involvement in infrastructure projects in
the United States” Albalate, D, Bel, G. and Geddes, R.
WP 2012/19 “Measuring Sovereign Bond Spillover in Europe and the Impact of Rating News” Claeys, P.
and Vašíþek, B.
WP 2012/18 “Entry Regulation Asymmetries and Gasoline Competition in a Mixed Motorway Network”
Albalate, D. and Perdiguero, J.
WP 2012/17 “Regulation of Port Charges in Spain: Global versus Local Competition” Fageda, X. and
Gonzalez-Aregall, M.
WP 2012/16 “Pass-through in dollarized countries: should Ecuador abandon the U.S. Dollar?” Marí del
Cristo, M.L. and Gómez-Puig, M.
WP 2012/15 “A Note on the Relationship Between the Cyclicality of Markups and Fiscal Policy” Claeys,
P. and Costa, L.F.
WP 2012/14 “Do intra- and inter-industry spillovers matter? CDM model estimates for Spain” Goya, E.;
Vayá, E and Suriñach, J.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 27
Research Institute of Applied Economics Working Paper 2013/09 pag. 27
27
WP 2012/13 “What Drives the Choice of Partners in R&D Cooperation? Heterogeneity across Sectors”
Badillo, E. and Moreno, R.
WP 2012/12 “Corruption and local politics: does it pay to be a crook?” Jiménez, J.L. and García, C.
WP 2012/11 “Not always sunny in paradise: prices and brand diversity in touristic areas supermarkets”
Campos, J.; Jiménez, J.L. and Suárez-Alemán, A.
WP 2012/10 “The institutional, economic and social determinants of local government transparency”
Albalate, D.
WP 2012/09 “The business excellence attraction composite index (BEACI). Design and apllication to the
municipalities of the Barcelona province” Murillo, J.; Romaní, J.; Suriñach, J.
WP 2012/08 “Policy options for the promotion of electric vehicles: a review” Perdiguero, J. and Jiménez,
J.L.
WP 2012/07 “Price differences between domestic and international air markets: an empirical application
to routes from Gran Canaria” Fageda, X.; Jiménez, J.L. and Díaz Santamaría, C.
WP 2012/06 “Building a “quality in work” index in Spain” López-Tamayo, J.; Royuela, V. and Suriñach, J.
WP 2012/05 “Mergers and difference-in-difference estimator: why firms do not increase prices?”
Jiménez, J.L. and Perdiguero, J.
WP 2012/04 “What attracts knowledge workers? The role of space, social connections, institutions, jobs
and amenities” Miguélez, E. and Moreno, R.
WP 2012/03 “What Drives the Urban Wage Premium? Evidence along the Wage Distribution” Matano, A.
and Naticchioni, P.
WP 2012/02 “Location Patterns of Creative Capital and Regional Disparities in Spain” Kerimoglu, E. and
Karahasan, B.C.
WP 2012/01 “The connection between distortion risk measures and ordered weighted averaging
operators” Belles-Sampera, J.; Merigó, J.M.; Guillén, M. and Santolino, M.
WP 2011/26 “Productivity and innovation spillovers: Micro evidence from Spain” Goya, E.; Vayá, E. and
Suriñach, J.
WP 2011/25 “The regional distribution of unemployment. What do micro-data tell us?” López-Bazo, E.
and Motellón, E.
WP 2011/24 “Vertical relations and local competition: an empirical approach” Perdiguero, J.
WP 2011/23 “Air services on thin routes: Regional versus low-cost airlines” Fageda, X. and Flores-Fillol,
R.
WP 2011/22 “Measuring early childhood health: a composite index comparing Colombian departments”
Osorio, A.M.; Bolancé, C. and Alcañiz, M.
WP 2011/21 “A relational approach to the geography of innovation: a typology of regions” Moreno, R.
and Miguélez, E.
WP 2011/20 “Does Rigidity of Prices Hide Collusion?” Jiménez, J.L and Perdiguero, J.
WP 2011/19 “Factors affecting hospital admission and recovery stay duration of in-patient motor victims
in Spain” Santolino, M.; Bolancé, C. and Alcañiz, M.
WP 2011/18 “Why do municipalities cooperate to provide local public services? An empirical analysis”
Bel, G.; Fageda, X. and Mur, M.
WP 2011/17 “The "farthest" need the best. Human capital composition and development-specific
economic growth” Manca, F.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 28
Research Institute of Applied Economics Working Paper 2013/09 pag. 28
28
WP 2011/16 “Causality and contagion in peripheral EMU public debt markets: a dynamic approach”
Gómez-Puig, M. and Sosvilla-Rivero, S.
WP 2011/15 “The influence of decision-maker effort and case complexity on appealed rulings subject to
multi-categorical selection” Santolino, M. and Söderberg, M.
WP 2011/14 “Agglomeration, Inequality and Economic Growth” Castells, D. and Royuela, V.
WP 2011/13 “A correlation sensitivity analysis of non-life underwriting risk in solvency capital
requirement estimation” Bermúdez, L.; Ferri, A. and Guillén, M.
WP 2011/12 “Assessing agglomeration economies in a spatial framework with endogenous regressors”
Artis, M.J.; Miguélez, E. and Moreno, R.
WP 2011/11 “Privatization, cooperation and costs of solid waste services in small towns” Bel, G; Fageda,
X. and Mur, M.
WP 2011/10 “Privatization and PPPS in transportation infrastructure: Network effects of increasing user
fees” Albalate, D. and Bel, G.
WP 2011/09 “Debating as a classroom tool for adapting learning outcomes to the European higher
education area” Jiménez, J.L.; Perdiguero, J. and Suárez, A.
WP 2011/08 “Influence of the claimant’s behavioural features on motor compensation outcomes” Ayuso,
M; Bermúdez L. and Santolino, M.
WP 2011/07 “Geography of talent and regional differences in Spain” Karahasan, B.C. and Kerimoglu E.
WP 2011/06 “How Important to a City Are Tourists and Daytrippers? The Economic Impact of Tourism
on The City of Barcelona” Murillo, J; Vayá, E; Romaní, J. and Suriñach, J.
WP 2011/05 “Singling out individual inventors from patent data” Miguélez,E. and Gómez-Miguélez, I.
WP 2011/04 “¿La sobreeducación de los padres afecta al rendimiento académico de sus hijos?” Nieto,
S; Ramos, R.
WP 2011/03 “The Transatlantic Productivity Gap: Is R&D the Main Culprit?” Ortega-Argilés, R.; Piva, M.;
and Vivarelli, M.
WP 2011/02 “The Spatial Distribution of Human Capital: Can It Really Be Explained by Regional
Differences in Market Access?” Karahasan, B.C. and López-Bazo, E
WP 2011/01 “I If you want me to stay, pay” . Claeys, P and Martire, F
WP 2010/16 “Infrastructure and nation building: The regulation and financing of network transportation
infrastructures in Spain (1720-2010)”Bel,G
WP 2010/15 “Fiscal policy and economic stability: does PIGS stand for Procyclicality In Government
Spending?” Maravalle, A ; Claeys, P.
WP 2010/14 “Economic and social convergence in Colombia” Royuela, V; Adolfo García, G.
WP 2010/13 “Symmetric or asymmetric gasoline prices? A meta-analysis approach” Perdiguero, J.
WP 2010/12 “Ownership, Incentives and Hospitals” Fageda,X and Fiz, E.
WP 2010/11 “Prediction of the economic cost of individual long-term care in the Spanish population”
Bolancé, C; Alemany, R ; and Guillén M
WP 2010/10 “On the Dynamics of Exports and FDI: The Spanish Internationalization Process” Martínez-
Martín J.
WP 2010/09 “Urban transport governance reform in Barcelona” Albalate, D ; Bel, G and Calzada, J.
WP 2010/08 “Cómo (no) adaptar una asignatura al EEES: Lecciones desde la experiencia comparada
en España” Florido C. ; Jiménez JL. and Perdiguero J.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 29
Research Institute of Applied Economics Working Paper 2013/09 pag. 29
29
WP 2010/07 “Price rivalry in airline markets: A study of a successful strategy of a network carrier against
a low-cost carrier” Fageda, X ; Jiménez J.L. ; Perdiguero , J.
WP 2010/06 “La reforma de la contratación en el mercado de trabajo: entre la flexibilidad y la seguridad”
Royuela V. and Manuel Sanchis M.
WP 2010/05 “Discrete distributions when modeling the disability severity score of motor victims”
Boucher, J and Santolino, M
WP 2010/04 “Does privatization spur regulation? Evidence from the regulatory reform of European
airports . Bel, G. and Fageda, X.”
WP 2010/03 “High-Speed Rail: Lessons for Policy Makers from Experiences Abroad”. Albalate, D ; and
Bel, G.”
WP 2010/02 “Speed limit laws in America: Economics, politics and geography”. Albalate, D ; and Bel, G.”
WP 2010/01 “Research Networks and Inventors’ Mobility as Drivers of Innovation: Evidence from
Europe” Miguélez, E. ; Moreno, R. ”
WP 2009/26 ”Social Preferences and Transport Policy: The case of US speed limits” Albalate, D.
WP 2009/25 ”Human Capital Spillovers Productivity and Regional Convergence in Spain” , Ramos, R ;
Artis, M.; Suriñach, J.
WP 2009/24 “Human Capital and Regional Wage Gaps” ,López-Bazo,E. Motellón E.
WP 2009/23 “Is Private Production of Public Services Cheaper than Public Production? A meta-
regression analysis of solid waste and water services” Bel, G.; Fageda, X.; Warner. M.E.
WP 2009/22 “Institutional Determinants of Military Spending” Bel, G., Elias-Moreno, F.
WP 2009/21 “Fiscal Regime Shifts in Portugal” Afonso, A., Claeys, P., Sousa, R.M.
WP 2009/20 “Health care utilization among immigrants and native-born populations in 11 European
countries. Results from the Survey of Health, Ageing and Retirement in Europe” Solé-Auró,
A., Guillén, M., Crimmins, E.M.
WP 2009/19 “La efectividad de las políticas activas de mercado de trabajo para luchar contra el paro. La
experiencia de Cataluña” Ramos, R., Suriñach, J., Artís, M.
WP 2009/18 “Is the Wage Curve Formal or Informal? Evidence for Colombia” Ramos, R., Duque, J.C.,
Suriñach, J.
WP 2009/17 “General Equilibrium Long-Run Determinants for Spanish FDI: A Spatial Panel Data
Approach” Martínez-Martín, J.
WP 2009/16 “Scientists on the move: tracing scientists’ mobility and its spatial distribution” Miguélez, E.;
Moreno, R.; Suriñach, J.
WP 2009/15 “The First Privatization Policy in a Democracy: Selling State-Owned Enterprises in 1948-
1950 Puerto Rico” Bel, G.
WP 2009/14 “Appropriate IPRs, Human Capital Composition and Economic Growth” Manca, F.
WP 2009/13 “Human Capital Composition and Economic Growth at a Regional Level” Manca, F.
WP 2009/12 “Technology Catching-up and the Role of Institutions” Manca, F.
WP 2009/11 “A missing spatial link in institutional quality” Claeys, P.; Manca, F.
WP 2009/10 “Tourism and Exports as a means of Growth” Cortés-Jiménez, I.; Pulina, M.; Riera i
Prunera, C.; Artís, M.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 30
Research Institute of Applied Economics Working Paper 2013/09 pag. 30
30
WP 2009/09 “Evidence on the role of ownership structure on firms' innovative performance” Ortega-
Argilés, R.; Moreno, R.
WP 2009/08 “¿Por qué se privatizan servicios en los municipios (pequeños)? Evidencia empírica sobre
residuos sólidos y agua” Bel, G.; Fageda, X.; Mur, M.
WP 2009/07 “Empirical analysis of solid management waste costs: Some evidence from Galicia, Spain”
Bel, G.; Fageda, X.
WP 2009/06 “Intercontinental fligths from European Airports: Towards hub concentration or not?” Bel, G.;
Fageda, X.
WP 2009/05 “Factors explaining urban transport systems in large European cities: A cross-sectional
approach” Albalate, D.; Bel, G.
WP 2009/04 “Regional economic growth and human capital: the role of overeducation” Ramos, R.;
Suriñach, J.; Artís, M.
WP 2009/03 “Regional heterogeneity in wage distributions. Evidence from Spain” Motellón, E.; López-
Bazo, E.; El-Attar, M.
WP 2009/02 “Modelling the disability severity score in motor insurance claims: an application to the
Spanish case” Santolino, M.; Boucher, J.P.
WP 2009/01 “Quality in work and aggregate productivity” Royuela, V.; Suriñach, J.
WP 2008/16 “Intermunicipal cooperation and privatization of solid waste services among small
municipalities in Spain” Bel, G.; Mur, M.
WP 2008/15 “Similar problems, different solutions: Comparing refuse collection in the Netherlands and
Spain” Bel, G.; Dijkgraaf, E.; Fageda, X.; Gradus, R.
WP 2008/14 “Determinants of the decision to appeal against motor bodily injury settlements awarded by
Spanish trial courts” Santolino, M
WP 2008/13 “Does social capital reinforce technological inputs in the creation of knowledge? Evidence
from the Spanish regions” Miguélez, E.; Moreno, R.; Artís, M.
WP 2008/12 “Testing the FTPL across government tiers” Claeys, P.; Ramos, R.; Suriñach, J.
WP 2008/11 “Internet Banking in Europe: a comparative analysis” Arnaboldi, F.; Claeys, P.
WP 2008/10 “Fiscal policy and interest rates: the role of financial and economic integration” Claeys, P.;
Moreno, R.; Suriñach, J.
WP 2008/09 “Health of Immigrants in European countries” Solé-Auró, A.; M.Crimmins, E.
WP 2008/08 “The Role of Firm Size in Training Provision Decisions: evidence from Spain” Castany, L.
WP 2008/07 “Forecasting the maximum compensation offer in the automobile BI claims negotiation
process” Ayuso, M.; Santolino, M.
WP 2008/06 “Prediction of individual automobile RBNS claim reserves in the context of Solvency II”
Ayuso, M.; Santolino, M.
WP 2008/05 “Panel Data Stochastic Convergence Analysis of the Mexican Regions” Carrion-i-Silvestre,
J.L.; German-Soto, V.
WP 2008/04 “Local privatization, intermunicipal cooperation, transaction costs and political interests:
Evidence from Spain” Bel, G.; Fageda, X.
WP 2008/03 “Choosing hybrid organizations for local services delivery: An empirical analysis of partial
privatization” Bel, G.; Fageda, X.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 31
Research Institute of Applied Economics Working Paper 2013/09 pag. 31
31
WP 2008/02 “Motorways, tolls and road safety. Evidence from European Panel Data” Albalate, D.; Bel,
G.
WP 2008/01 “Shaping urban traffic patterns through congestion charging: What factors drive success or
failure?” Albalate, D.; Bel, G.
WP 2007/19 “La distribución regional de la temporalidad en España. Análisis de sus determinantes”
Motellón, E.
WP 2007/18 “Regional returns to physical capital: are they conditioned by educational attainment?”
López-Bazo, E.; Moreno, R.
WP 2007/17 “Does human capital stimulate investment in physical capital? evidence from a cost system
framework” López-Bazo, E.; Moreno, R.
WP 2007/16 “Do innovation and human capital explain the productivity gap between small and large
firms?” Castany, L.; López-Bazo, E.; Moreno, R.
WP 2007/15 “Estimating the effects of fiscal policy under the budget constraint” Claeys, P.
WP 2007/14 “Fiscal sustainability across government tiers: an assessment of soft budget constraints”
Claeys, P.; Ramos, R.; Suriñach, J.
WP 2007/13 “The institutional vs. the academic definition of the quality of work life. What is the focus of
the European Commission?” Royuela, V.; López-Tamayo, J.; Suriñach, J.
WP 2007/12 “Cambios en la distribución salarial en españa, 1995-2002. Efectos a través del tipo de
contrato” Motellón, E.; López-Bazo, E.; El-Attar, M.
WP 2007/11 “EU-15 sovereign governments’ cost of borrowing after seven years of monetary union”
Gómez-Puig, M..
WP 2007/10 “Another Look at the Null of Stationary Real Exchange Rates: Panel Data with Structural
Breaks and Cross-section Dependence” Syed A. Basher; Carrion-i-Silvestre, J.L.
WP 2007/09 “Multicointegration, polynomial cointegration and I(2) cointegration with structural breaks.
An application to the sustainability of the US external deficit” Berenguer-Rico, V.; Carrion-i-
Silvestre, J.L.
WP 2007/08 “Has concentration evolved similarly in manufacturing and services? A sensitivity analysis”
Ruiz-Valenzuela, J.; Moreno-Serrano, R.; Vaya-Valcarce, E.
WP 2007/07 “Defining housing market areas using commuting and migration algorithms. Catalonia
(Spain) as an applied case study” Royuela, C.; Vargas, M.
WP 2007/06 “Regulating Concessions of Toll Motorways, An Empirical Study on Fixed vs. Variable Term
Contracts” Albalate, D.; Bel, G.
WP 2007/05 “Decomposing differences in total factor productivity across firm size” Castany, L.; Lopez-
Bazo, E.; Moreno, R.
WP 2007/04 “Privatization and Regulation of Toll Motorways in Europe” Albalate, D.; Bel, G.; Fageda, X.
WP 2007/03 “Is the influence of quality of life on urban growth non-stationary in space? A case study of
Barcelona” Royuela, V.; Moreno, R.; Vayá, E.
WP 2007/02 “Sustainability of EU fiscal policies. A panel test” Claeys, P.
WP 2007/01 “Research networks and scientific production in Economics: The recent spanish
experience” Duque, J.C.; Ramos, R.; Royuela, V.
WP 2006/10 “Term structure of interest rate. European financial integration” Fontanals-Albiol, H.; Ruiz-
Dotras, E.; Bolancé-Losilla, C.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 32
Research Institute of Applied Economics Working Paper 2013/09 pag. 32
32
WP 2006/09 “Patrones de publicación internacional (ssci) de los autores afiliados a universidades
españolas, en el ámbito económico-empresarial (1994-2004)” Suriñach, J.; Duque, J.C.;
Royuela, V.
WP 2006/08 “Supervised regionalization methods: A survey” Duque, J.C.; Ramos, R.; Suriñach, J.
WP 2006/07 “Against the mainstream: nazi privatization in 1930s germany” Bel, G.
WP 2006/06 “Economía Urbana y Calidad de Vida. Una revisión del estado del conocimiento en
España” Royuela, V.; Lambiri, D.; Biagi, B.
WP 2006/05 “Calculation of the variance in surveys of the economic climate” Alcañiz, M.; Costa, A.;
Guillén, M.; Luna, C.; Rovira, C.
WP 2006/04 “Time-varying effects when analysing customer lifetime duration: application to the
insurance market” Guillen, M.; Nielsen, J.P.; Scheike, T.; Perez-Marin, A.M.
WP 2006/03 “Lowering blood alcohol content levels to save lives the european experience” Albalate, D.
WP 2006/02 “An analysis of the determinants in economics and business publications by spanish
universities between 1994 and 2004” Ramos, R.; Royuela, V.; Suriñach, J.
WP 2006/01 “Job losses, outsourcing and relocation: empirical evidence using microdata” Artís, M.;
Ramos, R.; Suriñach, J.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 33
Research Institute of Applied Economics Working Paper 2013/09 pag. 33
33
doc_524182775.pdf
This paper analyses the determinants of broadband Internet access prices
in a group of 15 EU countries between 2008 and 2011. Using a rich panel
dataset of broadband plans, we show the positive effect of downstream
speed on prices, and report that cable and fibre-to-the-home technologies
are available at lower prices per Mbps than xDSL technology. Operators’
marketing strategies are also analysed as we show how much prices rise
when the broadband service is offered in a bundle with voice telephony
and/or television, and how much they fall when download volume caps
are included. The most insightful results of this study are provided by a
group of metrics that represent the situation of competition and entry
patterns in the broadband market. We show that consumer segmentation
positively affects prices. On the other hand, broadband prices are higher
in countries where entrants make greater use of bitstream access and lower
when they use more intensively direct access (local loop unbundling).
However, we do not find a significant effect of inter-platform competition
on prices.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2011/07 pàg. 1
Research Institute of Applied Economics Working Paper 2011/07 pag .1
1
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 33 pàg.
Research Institute of Applied Economics Working Paper 2013/09 33 pag.
“Broadband prices in the European Union: competition and
commercial strategies”
Joan Calzada and Fernando Martínez
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 2
Research Institute of Applied Economics Working Paper 2013/09 pag. 2
2
WEBSITE: www.ub.edu/irea/ • CONTACT: [email protected]
The Research Institute of Applied Economics (IREA) in Barcelona was founded in 2005, as a
research institute in applied economics. Three consolidated research groups make up the
institute: AQR, RISK and GiM, and a large number of members are involved in the Institute. IREA
focuses on four priority lines of investigation: (i) the quantitative study of regional and urban
economic activity and analysis of regional and local economic policies, (ii) study of public
economic activity in markets, particularly in the fields of empirical evaluation of privatization, the
regulation and competition in the markets of public services using state of industrial economy, (iii)
risk analysis in finance and insurance, and (iv) the development of micro and macro econometrics
applied for the analysis of economic activity, particularly for quantitative evaluation of public
policies.
IREA Working Papers often represent preliminary work and are circulated to encourage
discussion. Citation of such a paper should account for its provisional character. For that reason,
IREA Working Papers may not be reproduced or distributed without the written consent of the
author. A revised version may be available directly from the author.
Any opinions expressed here are those of the author(s) and not those of IREA. Research
published in this series may include views on policy, but the institute itself takes no institutional
policy positions.
1
Broadband prices in the European Union:
competition and commercial strategies
#
Joan Calzada
*
and Fernando Martínez
+
May 2013
Abstract
This paper analyses the determinants of broadband Internet access prices
in a group of 15 EU countries between 2008 and 2011. Using a rich panel
dataset of broadband plans, we show the positive effect of downstream
speed on prices, and report that cable and fibre-to-the-home technologies
are available at lower prices per Mbps than xDSL technology. Operators’
marketing strategies are also analysed as we show how much prices rise
when the broadband service is offered in a bundle with voice telephony
and/or television, and how much they fall when download volume caps
are included. The most insightful results of this study are provided by a
group of metrics that represent the situation of competition and entry
patterns in the broadband market. We show that consumer segmentation
positively affects prices. On the other hand, broadband prices are higher
in countries where entrants make greater use of bitstream access and lower
when they use more intensively direct access (local loop unbundling).
However, we do not find a significant effect of inter-platform competition
on prices.
Keywords: Telecommunications, Broadband prices, European Union,
Competition, Regulation
JEL: L51, L86, L96.
#
We thank Joan Ramon Borrell, Martin Cave, Xavier Fageda, Michal Grajek, Iñigo Herguera, Daniel
Montolio, Qmars Safikhani, Tommaso Valletti and Carlos Velasco for their useful comments and
suggestions. We also thank seminar participants at CMT (Barcelona), ITS (Viena), JEI (Murcia), Senatel
(Quito) and the Workshop on Telecommunications Economics (Barcelona). This research has received
financial support from the Spanish Ministry of Economics and Competitiveness (ECO2012-33392) and the
Regional Government of Catalonia (SGR2009-1066).
*
Corresponding author: Joan Calzada. Departament de Política Econòmica. Universitat de Barcelona.
Avinguda Diagonal 690. 08520 Barcelona. Tel. +34 934 034 743. Email: [email protected].
+
Fernando Martínez: Competition Commission and Universitat de Barcelona. Departament de Política
Econòmica. Avinguda Diagonal 690. 08520 Barcelona. Email: [email protected]
2
1. Introduction
Over the last decade, millions of people in the European Union have installed broadband
in their households
1
, enabling them to download information and use sophisticated digital
services
2
. Broadband Internet access is an essential component of inclusiveness in the 21
st
century, and households without access to this service are in risk of becoming marginalized
from society and economic opportunity. Several papers have analyzed the impact that
technological change and regulation have had in the expansion of Internet. However, little
attention has been given to how telecommunication operators have modified their business
strategies to respond to market evolution and competition. A detailed analysis of the
managerial behaviour of operators is essential if we are to understand the significant price
and quality differences that have emerged across Europe.
This paper examines the factors determining fixed broadband Internet prices in 15 EU
countries between 2008 and 2011
3
. We employ a rich data set that contains both the
commercial and technical characteristics of 2,204 plans offered by incumbent and entrant
operators. By using an instrumental variable approach we estimate a pricing equation using
three types of variable: (1) the technical characteristics of the plans; (2) the operators’
commercial strategies; and (3) the patterns of competition in the country. To the best of
our knowledge, this is the first paper to use information at the level of the individual firm’s
commercial plans to examine the determinants of prices in the EU.
Our study shows how operators adjust their prices to the technological characteristics of
the plans. First, we show that downstream speed has a positive and significant non-linear
impact on price: a 10 Mbps increase in the downstream speed increases the price by around
5%. And second, after controlling for downstream speed, we find that cable or fibre-to-
the-home (FTTH) broadband plans are cheaper than xDSL offers. This result is important
for orienting regulatory policy in the sector, as it questions the interest that operators might
have for deploying Next Generation Access Networks (NGAs).
This paper also shows the importance of the operators’ commercial practices. Thus while
flat rate plans are more expensive than metered plans (which are limited in volume), plans
that offer a broadband service bundled with voice telephony and/or television are
considerably higher in price, especially in the case of triple packages. Given the importance
of such bundling practices in many European markets today, we also analyse the main
determinants of this commercial practice.
Finally, we examine how levels of competition and regulation affect operators’ pricing
decisions. We show that incumbents set prices that are often significantly higher than those
of their competitors, which might be a consequence of such factors as their wider
1
The European Commission (2008) defines broadband internet access as “an access assuring an always-on
service with speeds in excess of 144 kbps. This speed is measured in download terms.”
2
During the nineties, broadband was delivered over cable and telephone lines. In the years that followed,
these technologies where upgraded and some operators began to deploy fibre to the home delivery, which
supports higher bandwidth (Greenstein and McDevitt, 2011a and 2011b).
3
In spite of their growing relevance, broadband mobile services are not included in this analysis. Note that
the commercial characteristics of mobile offers differ markedly from those of fixed broadband Internet
access. For example, download speed is significantly slower in the case of mobile offers (although new
wireless technologies such as LTE can provide speeds similar to fixed broadband)
3
coverage, reputation, or the existence of consumer switching costs. In contrast, the degree
of price dispersion and the overall number of plans offered in any given country have a
positive effect on prices. This result illustrates that market segmentation and consumer
confusion in the face of various marketing plans allow operators to increase their prices.
A further contribution of this paper is to identify the effects of access regulation. We find
that prices are higher in countries where entrants make much greater use of bitstream
access, and lower in markets where they rely more heavily on direct access (local loop
unbundling, LLU). Despite this, we observe no significant effect on prices when entrants
upgrade their own networks, nor do we find any effect of inter-platform competition
between xDSL, cable and FTTx. These results might be interpreted as a consequence of
the application of a “ladder of investment” approach (LOI),
4
whereby in order to promote
sector competition regulators need initially to facilitate the access of new entrants to the
incumbent’s network so as to guarantee service-based competition, and subsequently, once
these entrants have acquired experience and reputation they need to provide incentives to
entrants to invest in their own infrastructure. In this way, the long-term benefits of facility-
based competition can be reconciled with short-term price reductions.
The rest of the paper is organised as follows. Section 2 reviews the economic literature, so
as to highlight the contributions of this paper, and it describes the European broadband
market. Section 3 outlines our estimation strategy. Section 4 describes the data set. Section
5 presents the empirical strategy and results and Section 6 discusses the main contributions.
Finally, Section 7 concludes.
2. Literature review and European broadband market
2.1 Broadband access: the literature
The initial empirical literature on broadband Internet access focussed on the determinants
of its penetration. For example, Distaso, Lupi and Manenti (2006) reported the impact of
inter-platform competition on broadband penetration in 14 European countries from 2000
to 2004. They found that while inter-platform competition had a positive effect on
penetration, intra-platform did not play an important role. Other studies, including Höffler
(2007), highlighted the inefficiencies created by the duplication of existing platforms.
More recent papers have analysed the impact of access policies on the investment decisions
of firms and on the diffusion of the service.
5
Grajek and Röller (2012) examine the effects
of access regulation on incentives for investment in 20 countries in the period 1997-2006.
They report that regulation discourages the investment of incumbents and individual
entrants, and suggest that the European regulatory framework fails to provide incentives
for facility-based competition. They also examine the regulators’ response to infrastructure
investments, concluding that access regulation is not affected by entrants’ investments, but
rather that the intensity of regulation reflects the incumbent’s investments. Bouckaert, Van
4
The “ladder of investment” regulatory model was first identified by Cave (2006). See Cambini and Jiang
(2009) for an extensive review of the literature on this topic and Bourreau et al. (2010) for a critical analysis of
this regulatory approach.
5
A detailed review of the theoretical literature on access charges in telecommunications can be found in
Laffont and Tirole (2000), Arsmtrong (2002), and Vogelsang (2006).
4
Dijk and Verboven (2010) investigate the influence of competition on broadband
penetration in a sample of 20 OECD countries. They consider three entry patterns adopted
by broadband operators: (1) inter-platform competition, where the incumbent xDSL
operators compete with infrastructure-based operators (e.g. cable and FTTH); (2) facility-
based intra-platform competition, in which entrants lease some unbundled local loop
elements, but have to invest in their own equipment and facilities (e.g. LLU and share
lines); and (3) service-based intra-platform competition, where entrants resell the
incumbent’s services (bitstream access/resale). According to these authors, only
infrastructure-based competition increases the penetration of the service, while the other
types have little effect.
Few papers have undertaken specific country studies. Pereira and Ribeiro (2010) examine
the competition between xDSL and cable operators in Portugal. They find that inter-
platform competition (mainly between xDSL and cable) increases the diffusion of Internet
thanks to both the higher coverage of broadband access and the existence of lower prices.
More recently, Nardotto, Valletti and Verboven (2012) have analysed the impact of
unbundling on broadband penetration in the UK during the period 2005-2010 using micro
level information. They find that LLU had little or no effect on broadband penetration,
although it increased the quality of the service in terms of average broadband speed. On
the other hand, they show that inter-platform competition from cable increased local
broadband penetration.
Many of the above results contrast with those reported by Gruber and Koutroumpis (2013)
who, when drawing on data from 167 countries between 2000 and 2010, find that inter-
platform competition is an impediment to broadband adoption. They conclude that
markets that focus specifically on one type of technology typically present a more rapid
adoption process than that experienced in multi-technology markets. They suggest that this
is because the latter type of competition requires the duplication of networks, which
increases costs and, ultimately, prices. In contrast, full retail unbundling increases the
diffusion of the service.
While there have been several studies of broadband penetration in recent years, the
determinants of the prices of the service have received much less attention.
6
Explanations
for this include the absence of consistent data, and the fact that broadband services are
highly varied and typically offered jointly with voice telephony and television. An exception
is provided by Wallsten and Riso (2010), who examine a data set comprising 30 OECD
countries between 2007 and 2009. They find that downstream speed had a positive effect
on prices, that broadband plans with bit caps were on average offered at lower prices than
unlimited plans with contracts, and that plans with contracts were typically less expensive
than those without. Their paper has several points in common with our study, but they do
not examine the impact of access regulation on retail prices.
Greenstein and McDevitt (2011) also analyse the economic value created by the diffusion
of broadband Internet provided via xDSL and cable in the United States. They do not have
direct information on prices, but create a price index that allows them to adjust the price to
the progressive improvement in service quality. Taking this into account, they show that
broadband prices in the US fell slightly during the period 2004-2009. This is a very
6
Galperin (2012) describes the evolution of broadband prices in Latin America.
5
different evolution to that of the prices of electronic products, including laptops and
printers, where the quality-adjusted price falls have been significant.
2.2 The European broadband market
In July 2011, the average penetration level of fixed broadband Internet access in the
European Union was 27.2%
7
. However, there were significant differences across member
States. For example, while the penetration levels in Netherlands, Denmark and France were
39.3, 38.5 and 33.9%, respectively, in Romania, Bulgaria and Poland they were 14.6, 15.6
and 16.4%, respectively (Figure 1).
8
Figure 1: Fixed broadband penetration and incumbent’s penetration in 2011 (%)
Source: European Commission (2011a).
Figure 2: Fixed broadband prices in 2011 (€ PPP)
Least expensive offer (all ISPs): Basket 4096 kbps-8192 kbps, 5GB or 20 hours/month
Source: European Commission (2011b).
7
The penetration of large screen mobile broadband subscriptions (using dedicated data cards or USB
modems) was 7.5 % in July 2011.
8
See European Commission Implementation Reports and European Commission (2011a and 2011b).
4 5 5
7
11
10 10
9
11
8
12 11 10
12
18
14 14
12
23
17
23
15
9
12
15 14
24
17
14,6 15,6 16,4
17.0
19,3
20,8
20,9 21,4 21,6
22,1
22,3
23,6
23,9 24.0
25,2
25,6 26,6
27,2
29,2
30.0
31,4 31,6 32,4
32,4 32,7
33,9
38,5
39,3
RO BG PL SK LV EL PT HU LT CZ IT IE SI ES CY AT EE EU FI MT LU BE UK SE DE FR DK NL
Incumbent's penetration Total fixed broadband penetration
11 12 12
15
17 17
19
20
22 22 23 23 23
25 26 28
29
20
31
33 33 34
20
28
20
31
54
11
14
9
19
19
DK LT LV RO AT SE MT DE EL NL HU BG PL SK FI EE LU BE IE IT CZ SI FR UK ES PT CY
Tari ff Li ne rental
6
In recent years, the prices of fixed broadband Internet access have fallen significantly,
which is quite remarkable if we consider that operators have significantly improved the
quality of their offers. They often allow consumers to migrate at no cost to other offers
providing higher download speeds. Moreover, there is a trend towards bundling
complementary services, such as broadband with fixed voice and TV, and more recently
with mobile telephony (voice and data). Such packages allow operators to attract new
consumers (the bundle being cheaper than the sum of the single services) and, at the same
time, to gain the loyalty of their subscribers.
This situation has not prevented significant price differences across European countries
(Figure 2).
9
These differences can be accounted for in part by technical and commercial
features, but they also reflect the level of competition in the market. Thus, while in 2011
the incumbent’s market shares in Finland, Cyprus and Luxemburg were 79, 73 and 72%,
respectively; in the UK, Bulgaria and Romania they were 29, 29 and 30%. Many member
States have four or five alternative operators, but other markets are much more
fragmented. For example, in Germany there are around 100 regional entrants, though the
incumbent retains a 46% market share. The objective of our paper is to analyse how
technical, commercial and regulatory factors affect prices and to understand how they
determine price differences across member States.
Broadband access can be provided via several technologies. In the period we study, the
most frequently employed system is xDSL followed by cable modem, but some operators
use FTTH or wireless technologies such as 3G, Wimax and satellite. Around 77% of the
fixed access lines in European countries use the xDSL technology, which explains why the
average speed is still quite low (around 10 Mbps) and why there is more intra-platform than
inter-platform competition.
Incumbent operators are usually vertically integrated (except in Sweden, the UK and Italy)
and use xDSL (although some use cable, for example Portugal). Most entrants use the
incumbent’s network to provide their services and have to pay a regulated access fee. Cable
operators have built their own infrastructure, but they also need to sign interconnection
agreements with incumbent operators because of their limited national coverage.
In the EU, National Regulatory Authorities (NRAs) set interconnection prices in order to
guarantee an adequate development of competition. There are two mandatory types of
access. Entrants can access the incumbent’s network directly (LLU) or indirectly
(bitstream). At the same time, the direct access can be of three types: complete unbundling
of the local loop, where entrants pay to use the incumbent’s access lines without any
restriction; shared LLU, where entrants use the high frequencies of the access lines to
provide broadband and incumbents use the low frequencies to provide voice telephony;
and, shared LLU without voice telephony (naked ADSL), which is similar to the previous
service but voice telephony is offered over the Internet (VoIP). The main advantages of
unbundling are, therefore, to allow entrants to offer a differentiated service and to develop
their own commercial policy.
9
In the EU, retail prices of broadband services are not regulated. However, national regulators periodically
assess whether there is a “margin squeeze” that reduces the profitability of entrants. This might occur, for
example, when the level of interconnection prices (wholesale access prices) make impossible to entrants from
matching the incumbent’s offers.
7
In the case of indirect access (bitstream), entrants can access the incumbent’s network at
two levels: at the ATM level (or Gig-ADSL), where there are several geographical
interconnections, and at the IP level (or ADSL-IP), which is more expensive because there
are fewer interconnection points.
Price regulation of all these access services is an essential instrument for promoting
competition and investment. Regulated access prices determine in which part of the
incumbent’s network the entrants will choose to invest and influence both retail prices and
service quality. In the EU, following the “ladder of investment” (LOI) regulatory model,
regulators set the prices of indirect and direct access (bitstream and LLU) in order to
provide incentives to entrants to invest progressively in their own equipment. In spite of
this, the empirical literature is still unclear about the effectiveness of this strategy (Hazlett
and Bazelon, 2005; Bourreau and Dogan, 2006; Waverman et al., 2007; Bacache et al.,
2011, and Grajek and Röller, 2012). As Bourreau et al. (2010) explain, the main problem of
the LOI is that once entrants obtain some profits with bitstream access, their incentives to
invest may not be so high, creating a “replacement effect”. Moreover, the simultaneous
presence of multiple access levels can hinder incentives to access higher rungs on the
investment ladder. Our paper contributes to the empirical literature on access regulation by
assessing how the degree of intensity of each type of entry in a country affects retail prices.
3. Estimation strategy
This section examines the prices of broadband Internet access in 15 European countries in
the period 2008-2011. After adjusting for the hedonic features of the operators’ plans, we
analyse the impact of several commercial strategies adopted by operators, including
bundling and market segmentation. Additionally, we assess how different entry patterns
(bitstream, LLU and own network) induced by access regulations affect retail prices.
We estimate a model for the prices of broadband residential plans (p
mit
), where m is the
offer, i is the country, and t is the time period. The explanatory variables that we use in the
estimation can be grouped into three blocks: (1) technical characteristics of the service; (2)
the operators’ commercial strategies and (3) measures of competition and regulation in the
country. We also control for the degree of penetration of the service in the country, as well
as fixed country and time effects. Specifically, we estimate the following model:
pmit = a
0
+ a
1
Speedmit +a
2
Speedmit
2
+a
3
Technologymit
Technical characteristics
+a
4
Bundlingmit +a
5
VoIPmit +a
6
Datacapmit + a
7
Volumecapmit
Commercial characteristics
+a
8
Incumbentmit. + a
9
PDispersionit +a
10
NOffersit+a
11
HHIPlatit + a
12
Bitstreamit + a
13
Directaccessit + a
14
Ownetworkit
Measures of competition and regulation
+a
15
Penetrationit + a
16
Regiont +a
17
Timet +e
mit
8
The prices charged by operators in each country may vary according to the quality of the
service and the access technology. In our model, Speed is the downstream speed specified in
the technical details of the operators’ plans. To account for a possible non-linear
relationship between price and speed, the equation also includes the square of the speed.
Technology is the access technology used to provide the service; this might be xDSL, cable,
or fibre (FTTH). We expect each technology to have a different effect on the price given
that they require different levels of investment and bandwidths, and because consumers
might have different “perceptions” of the quality of each technology.
The price equation also includes commercial practices that may be adopted by operators.
Bundling refers to a broadband service provided together with voice telephony and/or
television. The commercialisation of broadband with other services implies additional costs
for the operators that can justify a price increase. For example, to be able to offer television
services, operators must first reach agreements with TV channels and pay them a
compensation. The variable VoIP reflects the situation where the broadband service is
bundled together with voice telephony but provided over IP, which reduces the operators’
costs (naked xDSL).
10
Datacap is a dummy variable that shows if the plan includes a restriction in the user’s
downstream volume (consumers have to pay an extra charge when they exceed the
established cap), and Volumecap measures the volume of data that the user can download. A
priori, we expect the capped offers to be cheaper than those with unlimited downstream
capacity, and also for prices to increase with the volume of the download limit.
We examine a group of variables that reflect the competition in the national markets.
Incumbent is a dummy that identifies the different pricing policies of the incumbent and the
entrants. Incumbents may have market power thanks to certain advantages of reputation or
to the existence of consumer switching costs. They may also have cost advantages over
their rivals. Yet, European operators may be the incumbent in one country and the entrant
in another or others. Thus, the costs of incumbent operators are also related to their
presence in several countries and to their bargaining power with their equipment providers.
HHIPlat is the concentration Herfindahl-Hirschman Index (HHI) in terms of technology
shares. A higher HHIPlat would mean a high concentration of a particular technology in a
given country. As discussed in Section 2, the empirical literature is ambiguous with regard
to the effect of inter-platform competition on the diffusion of the service (see, for example,
Bouckaert, van Dijk and Verboven, 2010 and Gruber and Koutroumpis, 2013), and its
effect on prices is also unclear. Here, a factor that should be considered is that inter-
platform competition allows operators to differentiate their services, which might offset
price reductions due to higher competition between different platforms.
PDispersion measures the price dispersion of all the plans commercialised in the country,
and NOffers is the number of plans offered by the incumbent. Both variables are proxies of
the degree of price discrimination in the market. Incumbents can offer a large number of
plans to better respond to competitors that target specific groups of consumers, but they
can also segment the market if they have market power.
10
Our model does not consider if the consumer can subscribe separately to each service (“menu à la carte”)
or if the client has no option other than to contract the bundle (so-called Tying). The reason is that it is
difficult to gather this information and also the problem of multicollinearity between the variables Tying and
Bundling.
9
Finally, a principal objective of this study is to determine how different types of entry in the
country (bitstream, LLU and the development of entrants’ own networks) affect the
operators’ pricing strategies. Bitstream, Directaccess, and Ownnetwork are explanatory variables
that reflect the relative importance of these entry patterns. Notice that important
distinctions can be drawn between Ownnetwork and HHIPlat. Thus, while the former
identifies an entrant that bypasses the incumbent’s network (implying the duplication of
networks), the latter reflects the presence of different technologies in the country, though
not necessarily the duplication of networks, i.e., the market might be geographically
segmented by technologies
11
. The inclusion of these variables in the pricing equation is
useful to analyse the effects of access regulation, although there are other important aspects
that also determine the entry strategy of operators, including the investment that is needed
to deploy the equipment, the operators’ perceptions of consumer willingness to pay for
high-quality services, or the regulatory institutions in the country.
Unfortunately, our data set does not provide any information about the number of clients
that subscribed to each of the commercial plans. Yet, the variable Penetration offers details
of the number of subscribers in each country for five different speed ranges. In the
presence of economies of scale, we expect operators to set lower prices as they have a
larger penetration and more subscribers to their plans. However, this effect may be
moderated when the increase in penetration is achieved as a result of extending service
coverage to high cost/low density regions.
4. The data
We use a panel data of residential retail broadband offers in 15 European member States
for the period December 2008 to June 2011. The countries analysed represent more than
80% of the total broadband access lines in the EU. On average the data set contains
around 550 offers per year and an overall total of 2,204 observations (Table 1). The sample
includes the tariffs of operators whose accumulative subscribers represent over 90% of the
broadband market in each country. Most of our data are drawn from Quantum-Web Ltd.
Data for the countries’ penetration rates and socio-economic variables are provided by the
European Commission, Eurostat and UNESCO.
Information concerning the technical and commercial characteristics of each offer is
collected primarily from the operators’ web sites by Quantum-Web. It should be stressed
that the prices offered on these sites might differ in some cases from those offered via
other sales channels (e.g. operators’ retail shops). Likewise, operators may negotiate
discounts with consumers when the latter have received better offers from rivals or when
they want to attract new subscribers. In the case of the technical characteristics, note that
the download speed advertised might not be uniform across all households.
The price of the offer used as the dependent variable is the monthly payment in euros
advertised by the operator and adjusted by each country’s purchasing power parity (PPP).
Some offers include the monthly line rental fee, which is commonly associated with voice
telephony. This is typically the case for xDSL, unless the voice service is provided via
VoIP, but not in the cases of cable and FTTH technologies. To account for this situation,
11
An example of market segmentation by technology is Belgium where the broadband lines in Flanders are
usually cable, while in Wallonia there is a more intensive use of xDSL.
10
we consider the price of each offer as being the sum of the access fee (the line rental) and
the broadband service tariff
12
. Note that in some cases the offer establishes limits to the
downstream volume, which means that consumers have to pay an additional charge once
this limit has been exceeded. Our price does not include these additional payments,
although the variables Datacap and Volumecap control for this situation.
We use two hedonic variables to estimate the price. Speed is the quality of the service
advertised on the operators’ web pages and is measured in Mbps. The minimum speed in
our sample is 0.128 kbps and the maximum is 500 Mbps. However, a significant number of
plans offer a quality between 10 and 30 Mbps (Table 2).
Source: Quantum Web-Ltd
The variable Technology reflects the type of access used to provide the service; this can be
xDSL, cable or FTTH. Taking this into account, the variable takes a value of 1, 2 or 3,
respectively. It should be stressed that the download speed is related to the technology
used in providing the service. Thus, xDSL cannot offer more than 30 Mbps, with the sole
exception of VDSL which can reach 50 Mbps. By contrast, cable supports speeds of up to
100 Mbps (DOCSIS3.0) and FTTH can attain download speeds of 1 Gbps. Also, an
operator’s bundling strategy depends on its technology. xDSL is usually bundled with voice
telephony, but not with television because good quality cannot be guaranteed. Cable and
FTTH, on the other hand, are often provided together with television.
Broadband access can be bundled with other services and commercialised at a single price.
To identify the incremental price that this implies, the variable Bundle takes values of 1, 2, 3
or 4, depending on whether broadband access is offered individually or bundled with fixed
voice telephony, TV, or with both
13
.
12
Prices do not include any other costs, such as the price of routers or the installation fee.
13
In contrast with Wallsten and Riso (2010), we have no information about the number of channels in triple
play packages.
Table 1: Descriptive statistics. Period 2008-2011
Variable Observations Mean Standard
Deviation
Minimum
Values
Maximum
Values
Price (euros) 2204 35.8 14.8 7.1 138.5
Price Single broadband (euros) 909 30.3 12.3 7.3 82.5
Price Broadband and voice (euros) 699 35.9 12.9 7.1 107.7
Price Broadband and TV (euros) 116 39.7 12.9 15.1 72.2
PriceBroadband, voice and TV (euros) 480 45.2 16.8 13.8 138.5
Price for Metered offers (euros) 410 36.7 14.7 7.1 79.8
Volume cap (Gb) 410 64.0 135.9 0.4 1000.0
Speed (Mbps) 2204 23.8 32.6 0.1 500.0
HHI inter-platform 2204 63.6 17.2 38.0 100.0
Bitstream access Index 2204 4.2 8.3 0.0 48.27
Direct access Index 2204 32.1 44.7 0.2 171.7
Own network access Index 2204 52.1 76.5 0.0 405.9
11
Datacap is a dummy variable that takes a value of 1 when the offer includes a limit to the
downstream volume, and 0 otherwise. The variable Volumecap measures the limit to the
download capacity (rated in GB) in offers with a downstream volume. Consumers usually
pay an extra charge if they exceed the downstream limit, but as explained above we do not
consider these charges in our analysis
14
.
Competition and regulation are essential aspects for understanding the operators’ pricing
policy. Our dataset provides information about the number of lines per operator in each
country, classified according to the technology used and the type of access (bitstream,
direct access, and own network). We use this information to construct concentration
indexes that measure the entry patterns in each country. Bitstream access concentration
(Bitstream) is defined as the square of the entrants’ bitstream market share (GigADSL o IP-
ADSL) in the country divided by the square of the incumbent’s market share. Direct access
concentration (Directaccess) is the square of the entrants’ direct access market share divided
by the square of the incumbent’s market share. Own network concentration (Ownnetwork) is
the sum of the squares of the entrants’ market share of own lines divided by the square of
the incumbent’s market share. As such, these indexes show the relevance of alternative
entry patterns in relation to the incumbent’s market share. In most European countries,
broadband services are in the main provided by legacy communication infrastructure,
where the incumbent operator maintains significant market power. Taking this into
account, we analyse how the intensity of different entry patterns with respect to the
incumbent’s position affects prices, and we also examine the response of prices to each
entry pattern.
A further instrument for measuring the competition level is the concentration of
technological platforms. To capture this, HHIPlat is the Herfindahl-Hirschman Index for
each country, which is estimated by adding the sum of the squares of market shares by
technology (xDSL, cable, FTTx).
PDispersion and NOffers can be interpreted as proxies of the degree of price discrimination,
and are calculated with the information obtained by Quantum-Web from the operators’
web sites. PDispersion is the standard deviation of broadband prices, and is calculated as the
square root of the variance in prices in each country for each year. NOffers is the number of
offers commercialised by the incumbent in each country and in each year. We might
alternatively have used the total number of offers commercialised in the country, but have
opted not to do so as we are unsure that our data set contains information about all
existing entrants.
Penetration is defined as the number of broadband subscriptions per 100 inhabitants in a
country. For the level of penetration we use EU information for five downstream speed
ranges: (1) below 2 Mbps, (2) 2-9.99 Mbps, (3) 10-29.99 Mbps, (4) 30-99.99 Mbps, and (5)
above 100 Mbps (ultrafast speed). The last two ranges are usually provided by cable or
FTTx, although the VDSL can also support speeds up to 50 Mbps.
14
Metered broadband plans charge for an additional capacity (GB) and some offers are time metered. The
extra costs (the so-called overage charges) are paid per GB or per a discrete number of extra GB, and less
frequently the plan charges per day, hour or minute above the cap limit. In some cases, no overage charges
are made and the broadband plans experience a sharp reduction in download speed once the cap has been
exceeded (bandwidth throttling).
12
The pricing equation also includes the dummy variable Region to account for the
unobserved heterogeneity in each market. The EU-15 countries have been clustered into
three groups owing to very high multicollinearity between country dummies. We have
tested that there are no significant differences in prices within each group after controlling
for the rest of the variables. In this way we avoid problems of multicollinearity and the loss
of degrees of freedom in the regression. The Region variable takes a value of 1 for Spain,
Belgium, Greece, Luxembourg and Portugal; 2 for Austria, Denmark, Netherlands, Finland
and Ireland; and 3 for France, Germany, Italy, Sweden and United Kingdom. Finally, Time,
is a dummy fixed effect for each year.
For illustrative purposes, Table 2 shows some of the most relevant characteristics of the
broadband plans for each country in 2011. The table highlights the differences across
countries in terms of price and download speed. Direct observation of these statistics
suggests that average prices might be affected by the speeds, but also by other factors such
as bundling or volume caps. The econometric analysis conducted in the next section seeks
to identify the main factors determining the operator prices and sheds some light on the
differences across countries.
Source: Quantum-Web Ltd
5. Estimation and results
This section presents an econometric multivariate analysis of the factors influencing
broadband Internet access prices. We estimate the pricing equation using three estimation
procedures: ordinary least squares (OLS), two-stage least squares (2SLS-IV) and three-stage
least squares estimations (3SLS-IV)
15
. We believe that our OLS estimates may suffer from
an endogeneity problem. Indeed, economic theory suggests that in countries where prices
are low, penetration levels might be higher. In such a case, the OLS coefficients of
15
We have discarded the use of a fixed effects model because of the severe attrition or identification of our
broadband plans (offers change over time or are substituted by others) and the little variance in our
observations during the period considered (with information covering four years). We have also ruled out the
use of a random effects model because the unobserved heterogeneity (the firm or country unobserved
characteristics) will be correlated with the explanatory variables in the pricing equation.
Table 2: Residential Broadband Plans. Characteristics by Country in 2011
Observa-
tions
Number of
Operators
Average
Price
(euros)
Average
Download
Speed
Bundling
(% bundled
plans)
Metered
Offers (%)
Average
Volume Cap
(Gb)
Bitstream
market
share (%)
ULL market
share (%)
Own Network
market share
(%)
Austria 71 7 39.6 29.9 58% 20% 58.4 2% 13% 25%
Belgium 32 5 42.6 20.2 34% 56% 25.5 5% 4% 37%
Denmark 12 3 25.9 27.3 58% 42% 208.3 7% 10% 22%
Finland 26 4 26.4 27.5 0% 4% 0.1 3% 2% 63%
France 46 5 34.3 52.9 93% - - 8% 43% 7%
Germany 55 10 26.6 28.2 65% 2% 1.8 7% 36% 12%
Greece 33 5 40.3 18.6 61% - - 2% 56% 0%
Ireland 38 4 38.1 17.3 55% 63% 29.2 20% 5% 26%
Italy 26 6 29.1 10.6 42% 15% 0.1 14% 30% 4%
Luxembourg 19 3 36.3 18.6 32% 16% 1.4 0% 11% 0%
Netherlands 60 9 39.3 29.6 53% - - 2% 13% 40%
Portugal 33 5 53.1 69.4 88% 39% 18.2 2% 9% 40%
Spain 47 7 56.9 20.9 89% 2% 0.0 6% 24% 18%
Sweden 53 5 28.7 36.9 32% - - 4% 14% 39%
UK 39 6 30.7 23.4 67% 31% 7.0 11% 38% 21%
13
penetration are likely to be biased downwards, and so we would erroneously conclude that
penetration has a smaller effect on price than it actually has. Taking this into account, we
use instrumental variable techniques and we examine the possibility of using different
socio-economic variables as instruments for Penetration: GDPpc is the gross domestic
product per capita; Density is the number of inhabitants in the country divided by its area in
square kilometres; Education is the percentage of population with upper-secondary
education; and Unemployment is the unemployment rate in each country. Data for GDP,
Density and Unemployment have been obtained from Eurostat, and Education from the
UNESCO data base.
We expect GDPpc, Density and Education to have a positive effect on the adoption of
internet and Unemployment to have a negative effect. We expect GDPpc to affect Internet
penetration but not broadband prices. In addition, both Price and GDPpc are adjusted by
each country’s PPP so as to account for differences in the cost of living across EU member
States. Density should reduce operator’s costs for network deployment and, hence, should
have a positive effect on coverage. However, the effect of Density on price is unclear, since
while Density is defined at a national level, population density can vary greatly across
regions, which might affect the pricing decisions of incumbents and entrants. As for
Education, we expect the more highly educated to be more likely to contract Internet, but
the level of education should have no impact on the operators’ price setting decisions.
Likewise, it should be noted that the concentration indexes (HHIPlat, Bitstream, Directaccess
and Ownnetwork) might also be affected by an endogeneity problem since the entrants’ entry
decisions could be determined simultaneously with prices. Yet, a high index value might
also reflect the presence of efficient entrants that are able to provide the service at a low
price, or who are perceived by consumers as being better. To account for this situation, our
model includes country fixed effects to represent the unobserved characteristics that
influence the efficiency of operators and, eventually, the level of retail prices. Examples of
these unobserved effects include investments, administrative constraints, and state aid
plans that are specific to each country.
Table 3 reports the OLS and 2SLS/3SLS-IV estimates of the pricing equation. The first
three columns present the OLS results. Specification 1 considers the technical
characteristics of the offers and the commercial strategies of the operators. Specification 2
includes the competition and regulatory variables and specification 3 adds Penetration.
Finally, specifications 4 and 5 show the results of the 2SLS and 3SLS estimations using as
instruments the socio-economic variables GDPpc, Density and Education. Since the
dependent variable price is included in logs, Penetration is interpreted as a semi-elasticity. All
specifications include the Time and Regional Effect dummy variables.
14
The estimates of the pricing equation are robust to the alternative specifications. Most of
the coefficients in the regressions are significant and their signs are in line with our
predictions.
16
As expected, Speed increases broadband prices. Specifically, a 10 Mbps
increase in speed raises broadband prices by an average of around 5%. However, the
negative sign of Speed
2
reveals that price exhibits diminishing returns with respect to speed.
Additionally, after controlling for other characteristics of the plans, xDSL appears to be
more expensive than cable and FTTH. Fibre and cable technologies can provide higher
speeds and better quality than xDSL, but it seems that this is not sufficient to enable
entrants to charge higher prices per Mbps than are charged for xDSL. Such a situation can
constitute an obstacle to the authorities’ objective of promoting investment in Next
Generation Access Networks (NGAs). This finding might be justified by the fact that
xDSL is often the only available technology in many regions and locations and as a
consequence incumbents fix uniformly higher prices for low speed offers per Mbps than
those set by cable and fibre operators, who tend to be present only in densely populated
areas where there are several competitors.
As for the operators’ commercial strategies, unlimited plans (flat rates) charge higher prices
than metered plans. In the case of metered offers, when the volume cap is exhausted the
average increase in the price for an extra GB is slightly lower than 0.05%. On the other
hand, bundles of services are more expensive than stand-alone broadband plans. Offers
combining broadband access and TV are more expensive than those that bundle
broadband access and voice telephony. Specifically, broadband packages with voice or TV
are 15 and 18% more expensive, respectively. When the offer includes both voice
16
We test for multicollinearity using the variance inflator factor (VIF). Its mean equals 1.9 and the two highest
values are 4.9 for Speed and 3.4 for Speed
2
; the rest of the VIF values are below 3.
Table 3: Estimation Results: All Broadband plans. OLS; 2SLS/3SLS-IV
Dependent variable Specification 1 Specification 2 Specification 3 Specification 4 Specification 5
Log Price (price) OLS OLS OLS 2SLS-IV 3SLS-IV
Independent variables Coefficient Coefficient Coefficient Coefficient Coefficient
Penetration -0.0057* (0.00312) -0.017** (0.00723) -0.019** (0.00723)
Speed 0.0065*** (0.00146) 0.0069*** (0.0014) 0.006*** (0.00159) 0.0042*** (0.00185) 0.0047*** (0.00107)
Speed² -1.1e-05** (0.000004 -1.1e-05** (0.000004 -9.2e-06* (0.000004) -5.4e-06* (0.000005) -1.5e-05* (0.000003)
Technology dummy (reference xDSL)
Cable -0.092* (0.05150) -0.11** (0.04710) -0.11** (0.04577) -0.15*** (0.03449) -0.10** (0.02133)
FTTx -0.11** (0.0582) -0.11** (0.05026) -0.12** (0.05135) -0.14** (0.04405) -0.11** (0.02515)
Metered offer (Data cap) -0.098 (0.05860) -0.17*** (0.05739) -0.18*** (0.05816) -0.18*** (0.05773) -0.16*** (0.01878)
Volume cap 0.00043 (0.00030) 0.00049* (0.00023) 0.00049** (0.00022) 0.00048** (0.00021) 0.00042** (0.0001)
Bundle (refrence stand alone broadband)
Internet and voice 0.18*** (0.02848) 0.17*** (0.02719) 0.17*** (0.02591) 0.17*** (0.02372) 0.15*** (0.01482)
Internet and tv 0.19*** (0.04024) 0.20*** (0.03544) 0.20*** (0.03318) 0.21*** (0.03008) 0.18*** (0.02804)
Internet, voice and tv 0.36*** (0.05198) 0.34*** (0.04016) 0.34*** (0.03795) 0.34*** (0.03351) 0.31*** (0.01743)
VoIP -0.073 (0.04870) -0.077 (0.04668) -0.072 (0.04296) -0.063 (0.03831) -0.078*** (0.0232)
Incumbent 0.15*** (0.03165) 0.14*** (0.03282) 0.14*** (0.03265) 0.12*** (0.03099) 0.15*** (0.01836)
HHIPlat 0.0014 (0.00110) 0.00089 (0.00116) -0.00016 (0.00080) -0.00035 (0.00113)
Bitstream 1.1*** (0.25199) 1*** (0.23629) 0.96*** (0.21353) 0.99*** (0.11317)
Direct Access -0.16*** (0.05061) -0.14** (0.04868) -0.092* (0.04770) -0.073** (0.03442)
Ownnetwork 0.004 (0.01470) -0.00062 (0.01290) -0.0095 (0.01254) -0.0099 (0.01112)
PDispersion 0.022*** (0.00710) 0.02*** (0.00625) 0.017*** (0.00530) 0.02*** (0.00327)
Noffers Incumbent 0.0023 (0.00222) 0.0043* (0.00225) 0.0081** (0.00409) 0.0026 (0.00289)
R
2
0.48 0.52 0.53 0.51 0.49
Number of observations (N) 2204 2204 2204 2204 2204
Note: Robus t standard errors are i n parenthes i s . Si gni fi cance at * 10%, ** 5%, *** 1% l evel s .
15
telephony and television, prices rise by about 30%. By contrast, offers that include
broadband and voice can be up to 7% cheaper if the voice service is supplied over IP
(VoIP).
In the case of the competition variables, incumbents’ plans are around 15% more
expensive than entrants’ offers, which might reflect the formers’ dominant position in the
market and “unobserved” quality differences as perceived by consumers. PDispersion and
Noffers exhibit a positive effect on prices, which suggests that when firms are better able to
screen consumers they set higher prices, and that a large number of plans of varying
characteristics and prices might generate a certain confusion among consumers, thus,
allowing firms to set higher prices (Hoernig, 2001). We also find that technological
concentration, measured with the variable HHIPlat, does not have a significant effect on
prices.
Our estimates show the entry pattern to be an important determinant of broadband prices
in a country. While the intensity of Bitstream entry increases prices, Direct Access (LLU)
reduces them, and Ownetwork does not entail a significant effect. On the other hand, the
coefficient associated with Bitstream is higher than that associated with Direct Access,
indicating that with an equivalent variation in the index there will be a greater price reaction
with Bitstream. One explanation is that LLU allows operators to differentiate their products
and to develop their own commercial strategies, which may avoid large price reductions for
equivalent levels of entry.
As expected, the OLS estimate shows that the Penetration variable has a negative effect on
prices. However, OLS may produce biased estimators if price and penetration are
determined simultaneously. To account for this, we applied the instrumental variable
approach (2SLS) using as our instrument a combination of the previously defined socio-
economic variables GDPpc, Density, Education and Unemployment.
Table 4 presents the result of the Hausman test that confirms that Penetration is an
endogenous variable. It also shows the results of the validity tests for different
combinations of instruments. The instruments in the second and third specification pass
the Hansen’s J test for over-identifying restrictions and are also shown to be suitable for
explaining Penetration. However, the validity test results for GDPpc and Unemployment (Test
Result 1) are not satisfactory. Moreover, Hansen’s J test value for the third specification is
lower than the second value. Therefore we use GDPpc, Density and Education as our
instruments.
Finally, we apply the instrument suitability tests (the F-statistic in the first stage regression
of the variable Penetration) to verify that the instruments are strong. Additionally, the
computed standard errors are robust to any bias from heteroskedasticity and they are also
clustered according to observations from the same country.
Table 4: Endogeneity test for Penetration
Hausman endogeneity test. Ho: Penetration exogenous p-value= 0.0000
Instruments for Penetration GDPpc, Unemployment
GDPpc, Education,
Density
GDPpc, Education,
Density, Unemployment
Test Test Result 1 Test Result 2 Test Result 3
Hansen J test. Ho: instruments exogenous p-value=0.1162 p-value=0.1691 p-value=0.1140
Validity of Instruments Ho: weak instruments p-value=0.1489 p-value=0.0001 p-value=0.0001
16
Usually, in simultaneous equation models (SEM), the variables determined simultaneously
(in this case, Price and Penetration) are correlated with the error term in the regression,
causing the OLS coefficients to be biased and inconsistent. To overcome this problem, we
have also estimated the pricing equation using 3SLS, which is a more efficient method
17
.
The results of the 2SLS and 3SLS estimations are shown in columns 4 and 5 in Table 3. We
find that the technical, commercial, and competition coefficients are robust to the use of
these estimation techniques. In the case of Penetration the sign of the coefficient is
maintained, but its value is higher than that obtained with the OLS regression. In the case
of the 3SLS-IV regression, the coefficient ??
P-3SLS
= -0.019 implies that a one-percentage-
point increase in the penetration level is followed by a 1.9% fall in price. The negative
estimate of Penetration instrumented through the socio-economic variables more than triples
the OLS coefficient, showing that the OLS Penetration coefficient was downward biased.
We also considered the effect of using the lagged Penetration variable as our instrument,
where LagPenetration and Lag2Penetration are defined as the variable lagged one and two
periods, respectively. Table 5 shows the Hausman endogeneity test for the lags of
Penetration. Both variables are endogenous but their Hausman chi-square test values are
lower than that for Penetration. The latter can be accounted for by the fact that prices and
the subscription decision are determined in the same period and, as such, the simultaneity
problem should be reduced for lagged periods of penetration. We ran the same OLS
regression as specification 3 in Table 3 and found that the coefficient is ??
LagPenetration-OLS
= -
0.007 when we consider one lag and ??
Lag2Penetration-OLS
= -0.0076 when we consider two lags of
Penetration, and in the two cases the coefficients are significant at the 5% level. These results
confirm that the simultaneity bias is downwards.
6. Discussion
Below we discuss in greater detail two key features that affect the operators’ commercial
policies: their bundling strategies and entry patterns.
Bundling strategies - A commercial strategy widely adopted by telecom operators in the
EU is that of bundling several services together in the same offer. Our analysis shows how
each additional service in such a bundle increases the price of the broadband plan.
However, standalone broadband and bundled offers can be considered as different
products and operators price them differently. To account for this possibility, we present
17
The 3SLS estimation of the model takes into account the structural equations for both Price and Penetration
and uses the cross-equation correlation of errors to provide more efficient estimators than the 2SLS
estimation. The procedure involves an additional step to the 2SLS procedure: namely, it uses the residuals
from both the Price and Penetration structural equations to construct a covariance matrix of errors that
provides a more efficient estimation.
Table 5: Endogeneity test for Penetration and Lags of Penetration
Hausman endogeneity test. Ho: Penetration exogenous Chi2(1)= 38.345 (p-value = 0.0000)
Hausman endogeneity test. Ho: LagPenetration exogenous Chi2(1)= 31.801 (p-value = 0.0000)
Hausman endogeneity test. Ho: Lag2Penetration exogenous Chi2(1)= 7.725 (p-value = 0.0054)
17
separate estimations for standalone broadband and for the bundle of broadband and voice
services
18
. Table 6 shows that these new estimations are very similar to those found when
we consider all offers. In the 2SLS and 3SLS estimations, Penetration is still instrumented by
the socio-economic variables GDPpc, Density, and Education, but its coefficient is no longer
significant
19
. By contrast, HHIPlat is now significant, indicating that a higher concentration
of one technological platform (less inter-platform competition) raises prices. In the case of
bundles, this result might be further reinforced because cable and FTTH are more likely to
offer TV than is the case with ADSL. As a result, technological concentration generates
higher prices for bundled offers.
18
In the case of bundle offers, we are unable to identify the price of each particular service, but we can
observe how the price of an offer increases as other services are integrated. Wallsten and Riso (2008) also
adopt this approach when analyzing bundling.
19
The penetration information we use is based on the whole sample since it is not possible to distinguish
between penetrations rates that depend on bundled plans, on the one hand, and those that depend on
unbundled plans, on the other.
Table 6: Estimation Results: Standalone Broadband and Bundles. OLS; 2SLS/3SLS-IV
Dependent variable
Log Price (price) OLS 2SLS-IV 3SLS-IV OLS 2SLS-IV 3SLS-IV
Independent variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient
Penetration -0.0019 0.0036 0.0038 -0.0038 -0.0041 -0.0057
(0.00331) (0.00824) (0.00919) (0.00229) (0.00831) (0.00899)
Speed 0.0088*** 0.0096*** 0.01*** 0.0071*** 0.0070*** 0.0075***
(0.00185) (0.00160) (0.00144) (0.00185) (0.00239) (0.00215)
Speed² -1.5e-05* -1.6e-05* -1.9e-05*** -1.7e-05* -1.6e-05* -2.3e-05***
(0.000005) (0.000005) (0.000003) (0.000008) (0.000008) (0.000008)
Technology dummy (reference xDSL)
Cable -0.18*** -0.17*** -0.12*** -0.11** -0.11** -0.08***
(0.05689) (0.05295) (0.02962) (0.05119) (0.04786) (0.02969)
FTTx -0.17** -0.15** -0.09** -0.086 -0.087* -0.057**
(0.07424) (0.06814) (0.04334) (0.06263) (0.05163) (0.03739)
Metered offer (Data cap) -0.15* -0.14** -0.1*** -0.2** -0.2*** -0.18***
(0.07203) (0.07050) (0.03117) (0.07460) (0.07306) (0.02791)
Volume cap 0.0003 0.0003 0.00022 0.0006** 0.0006*** 0.00055***
(0.00030) (0.00029) (0.00014) (0.00023) (0.00021) (0.00015)
VoIP -0.038 -0.038 -0.043
(0.03779) (0.03572) (0.02913)
Incumbent 0.11** 0.12** 0.14*** 0.18*** 0.18*** 0.2***
(0.04379) (0.04662) (0.02936) (0.04249) (0.03574) (0.02773)
HHIPlat 0.0016* 0.0021*** 0.0024** 0.0042* 0.0042* 0.004**
(0.00077) (0.00066) (0.00113) (0.00217) (0.00218) (0.00131)
Bitstream 0.89*** 0.93*** 0.96*** 0.86*** 0.86*** 0.86***
(0.27161) (0.22084) (0.18773) (0.21209) (0.20025) (0.14252)
Direct Access -0.1 -0.12** -0.11** -0.17*** -0.17*** -0.16***
(0.06613) (0.05136) (0.04758) (0.03608) (0.03444) (0.04247)
Ownnetwork -0.0058 0.00057 0.0036 -0.019 0.02 0.022
(0.01037) (0.01428) (0.01491) (0.08236) (0.08320) (0.05283)
PDispersion 0.023*** 0.026*** 0.029*** 0.022*** 0.022*** 0.022***
(0.00649) (0.00711) (0.00568) (0.00693) (0.00734) (0.00458)
Noffers Incumbent -0.00044 -0.0028 -0.0094* 0.0075* 0.0075* 0.0052*
(0.00383) (0.00372) (0.00487) (0.00412) (0.00415) (0.00307)
Regional Effect (reference Region 1)
Region 2 -0.21*** -0.22*** -0.21*** -0.24*** -0.24*** -0.24***
Region 3 -0.41*** -0.41*** -0.39*** -0.42*** -0.42*** -0.42***
Time dummy (reference t=2008)
t=2009 0.071 0.069 0.064** 0.061 0.061 0.059**
t=2010 0.081 0.079* 0.072** 0.12** 0.12** 0.12***
t=2011 0.11** 0.11** 0.097*** 0.15** 0.15** 0.15***
Cosntant 3.1*** 2.9*** 2.9*** 3*** 3.1*** 3.1***
R
2
0.42 0.42 0.40 0.58 0.58 0.57
Number of observations (N) 909 909 909 699 699 699
Note: Robust standard errors are i n parenthesi s. Si gni fi cance at * 10%, ** 5%, *** 1% l evel .
Stand-alone Broadband Bundle: Broadband + Fixed Voice Telephony
18
We are also interested in identifying the factors that serve as incentives for operators to
commercialise bundled services, since significant differences can be observed across
European countries (Table 2). In order to address this question we run logit and probit
discrete choice models that consider 1,336 observations collected on a yearly basis between
2008 and 2011. This regression focuses only on xDSL offers since bundling decisions differ
substantially for cable and FTTH. For this sample, 42% of the offers are broadband
standalone offers, 37% are bundles of broadband and voice, and 21% include TV,
broadband and voice.
The results of the estimation for Bundle are presented in Table 7. As expected, Speed
increases the probability of broadband being bundled with other services. Clearly, high
download speeds are necessary in order to offer TV. However, no significant distinction is
observed between the bundling strategies of incumbents and entrants. Market and
technological concentration (HHIOper and HHIPlat) are negatively associated with the
probability of bundling. This suggests that bundling is used by operators to attract
consumers in competitive markets, whereas in more concentrated markets operators prefer
to sell additional services separately. We also consider the effects of price dispersion and
find that it increases the probability of bundling. Therefore, in markets with a wide-range
of prices, operators might commercialise bundles of services to differentiate their plans.
Finally, no correlation is found between bitstream and direct access and bundling, although
bundling is found to be less likely in markets where more entrants bypass the incumbent’s
network (own network).
Table 7: Logit/Probit for Bundling
Bundling Logit Probit
Independent variables Coefficient Coefficient
Speed 0.033*** (0.00827) 0.02*** (0.004516)
Incumbent 0.38 (0.2556) 0.23 (0.15672)
HHIOperator -0.093** (0.03781) -0.057** (0.02286)
HHIPlat -0.035** (0.01614) -0.022** (0.00998)
PDispersion 0.062* (0.03288) 0.038* (0.01957)
Bitstream -0.9 (0.7889) -0.56 (0.4786)
Direct Access 0.1 (0.4771) 0.075 (0.2862)
Own network -2.8*** (0.8134) -1.7*** (0.5001)
GDPpc 0.003 (0.0053) 0.0022 (0.003317)
Education -0.017 (0.01455) -0.011 (0.00888)
Density 0.00072 (0.00059) 0.00045 (0.00037)
Regional Effect (ref: Region 1)
Region 2 -0.3 (0.3644) -0.19 (0.2239)
Region 3 0.22 (0.4141) 0.14 (0.2555)
Constant 5.8* (2.3438) 3.6** (1.4299)
Pseudo R
2
0.11 0.11
Number of observations (N) 1336 1336
Dependent variable
Note: Robust standard errors are i n parenthesi s.
Si gni fi cance at * 10%, ** 5%, *** 1% l evel .
19
Entry pattern - One of the main results of our analysis is that broadband prices are
negatively related with LLU entry intensity, and positively related with bitstream entry
intensity. This result implies that during the period analysed intra-platform facility-based-
competition was more effective in reducing prices than was intra-platform service-based-
competition, once we control for other characteristics of the offer. At this juncture, it
should be stressed that the intensity of each country’s entry pattern is influenced by the
regulation of access charges. In fact, European regulators usually follow the LOI approach,
which proposes setting higher prices for bitstream so as to induce entrants to use direct
access (Cave, 2006; Höffler, 2007, Bourreau, Dögan and Manant, 2010). As shown in
Bacache, Bourreau and Gaudin (2011), this regulatory strategy may have facilitated the
migration from bitstream access lines to LLU, although their study does not find any ladder
effect between LLU and entrants deploying their own network.
The regulation of access charges in the EU has been designed to provide incentives for
investment by entrants; however, in this paper we have also shown that it affects retail
prices. Prices are higher in countries where entrants make greater use of bitstream entry
and lower in countries where they make greater use of direct access. In fact, while the
establishment of high access charges for bitstream might act as an incentive to incumbent
operators to set higher uniform prices for their Internet plans, the establishment of low
access charges for direct access might not result in an equivalent response by operators,
since this type of entry requires major investment and because it allows operators to
differentiate their product (Nardotto, Valletti and Verboven, 2012). This situation is
reflected in our estimates of the Bitstream and Directaccess concentration indexes, since the
value of the coefficient is much greater for the former than for the latter. For instance, in
specification 5, ??
Bitstream-3SLS
=0.99 and ??
Directaccess-3SLS
=-0.073, which reflects the greater increase
in prices due to bitstream concentration than the relative decrease due to LLU
concentration.
Our results also suggest that inter-platform competition does not affect significantly retail
prices. Entrants that invest in cable and FTTH networks are able to commercialise high-
quality services, but this does not necessarily mean a reduction in their retail prices.
Moreover, although cable and FTTH plans involve lower prices per Mbps than those
charged by xDSL plans, the former typically offer more downstream speed and additional
services such as TV, which increase prices. A further important aspect that should be taken
into account when interpreting our results is that in our model competition is introduced at
the national level, but in some countries technologies are geographically segmented and so
there is little competition between them.
6. Conclusions
This paper has analysed the determinants of the prices of broadband Internet access in 15
countries of the EU between 2008 and 2011. Our econometric model has focused on three
types of variable: (1) the technical characteristics of the offers; (2) the operators’
commercial strategies; and (3) the regulation and competition in the country. Besides, we
have controlled for the potential endogeneity of broadband penetration by using the
instrumental variable approach (2SLS/3SLS) and employing as instruments a group of
socio-economic variables that determine the demand for broadband services in each
country.
20
Our analysis has revealed that downstream speed is a significant driver of price: a 10 Mbps
increase in the download speed causes prices to rise by an average of around 5%.
Additionally, we have found that the price per Mbps paid for the service via cable or fibre
technologies is lower than that provided by xDSL, although these technologies usually
provide higher download speeds. On the other hand, the xDSL service usually involves
more intra-platform competition and has a wider coverage than other technologies. In this
context, an important question that emerges is whether consumer willingness to pay for
fibre and cable plans is sufficiently high to encourage operators to invest in NGAs.
The operators’ marketing strategies also play an important role in determining prices. When
the broadband service is bundled with voice telephony, the price increases by around 15%
and when it is bundled with both voice telephony and television it increases by around
30%. By contrast, consumers that contract the voice service through VoIP obtain
significant price reductions. We have also shown that operators are less likely to bundle
their offers when the market is less competitive or when there is less inter-platform
competition.
This paper has also contributed to the literature that analyzes how regulatory policy can
influence the development of the broadband market. Here, we have shown that broadband
prices are higher in countries where entrants make greater use of bitstream access and
lower in those countries making greater use of LLU. We have found no evidence that inter-
platform competition and stand-alone entry (the last rung on the “ladder of investment”
approach) reduce prices. This means that when entrants bypass the incumbents’ networks,
retail prices are not reduced. This might be because this type of entry does not generate
sufficient competition, and because the operators offer high-quality products that are more
expensive. All in all, our results can be interpreted as an assessment of the “ladder of
investment” approach, and they show the benefits of facilitating migration from bitstream
to LLU entry.
One limitation of our study is that we have not considered mobile broadband plans offered
via smartphones or dongles. Mobile broadband demand is boosting and future research
needs to consider the impact of the firms’ commercial strategies for mobile broadband
offers. For example, a rising number of operators are currently offering bundled mobile
and fixed broadband packages, which minimises the consumer churn rate (Prince and
Greenstein, 2011).
21
7. References
Andrés, L., D. Cuberes, M. Diouf and T. Serebrisky (2010), “The diffusion of the Internet: A cross-
country analysis, Telecommunications Policy, 34, 323-340.
Armstrong, M. (2002), “The theory of access pricing and interconnection”, in M. Cave, S.
Majumdar, and I. Vogelsang (eds.), Handbook of Telecommunications Economics, Vol. 1. Amsterdam:
Elsevier.
Bouckaert, J., T. Van Dijk and F. Verboven (2010), “Access regulation, competition, and
broadband penetration: An international study”, Telecommunications Policy, 34, 661-671.
Bacache, M., M. Bourreau and G. Gaudin (2011), “Dynamic Entry and Investment in New
Infrastructures: Empirical Evidence from the Telecoms Industry”, Working Papers in Economics and
Social Sciences, 11-01.
Bohlin, E. and O. Teppayayon (2010), “Functional separation in Swedish broadband market: Next
step of improving competition”, Telecommunications Policy, 34, 375-383.
Bourreau, M. and P. Dögan (2006), “Build or Buy Strategies in the Local Loop”, American Economic
Review, Papers and Proceedings, 96 (2), 72-76
Bourreau, M., P. Dögan and M. Manant (2010), “A critical review of the ladder of investment
approach”, Telecommunications Policy, 34 (11), 683-696.
Cambini, C. and Y. Jiang (2009), “Broadband Investment and Regulation: A literature Review,”
Telecommunications Policy, 33 (10-11), 559-574.
Cava, I. and A. Alabau (2006), “Broadband policy assessment: a cross national empirical analysis”,
Telecommunications Policy, 30, 445-463.
Cave, M. (2006), “Encouraging infrastructure competition via the ladder of investment”,
Telecommunications Policy, 30, 223–237.
Czernich, N., O. Falck, T. Kretschmer and L. Woessmann (2011), “Broadband Infrasctructure and
Economic Growth”, The Economic Journal, 121 (May), 505-532.
Cincera, M., L. Dewulf and A. Estache (2012), “On the (In)Effectiveness of Policies to Promote
Broadband Diffusion in Europe (2003-2010): An Economic Assessment, Ecore Discussion Paper.
Deligiorgi C., C.Michalakelis A., Vavoulas and D. Varoutas (2008), “Nonparametric estimation of a
hedonic price index for ADSL connections in the European market using the Akaike Information
Criterion”, Telecommunication System (2007) 36: pp173-179.
Distaso, W., P. Lupi and G. Manenti (2006), “Platform Competition and Broadband Uptake:
Theory and Empirical Evidence”, Information Economics and Policy, 18 (1), 87-106.
European Commission (2009).“15th Progress Report on the Single European
Electronic Communications Market – 2009“.
European Commission (2011a), “Broadband access in the EU: situation at 1 July 2011, Information
Society and Media Directorate-General 2009, COCOM09-29 Final.
22
European Commission (2011b), “Broadband Internet Access Costs (BIAC)”. Final Report,
Information Society and Media Directorate-General.Van Dijk – Management Consulting.
Galperin, H. (2012), “Precios y calidad de la banda ancha en América Latina: Benchmarking y
tendencias”, Documentos de Trabajo, Universidad de San Andrés.
Grajek, M., and L.-H. Röller (2012), “Regulation and investment in network industries: Evidence
from European telecoms”, Journal of Law and Economics, 55 (1): 189-216.
Greenstein, S. and R. McDevitt (2011), “The Broadband Bonus: Estimating broadband Internet’s
economic value, Telecommunications Policy, 35 (7) , 617-632.
Greenstein, S. and R. McDevitt (2011), “Evidence of a modest price decline in US broadband
services”, Information, Economics and Policy, 23 (2), 200-211.
Greenstein, S. and J. Prince (2011), “Does service Bundling reduce churn”, mimeo.
Gruber, H. and P. Koutroumpis (2013) Competition enhancing regulation and diffusion of
innovation: the case of broadband networks, Journal of Regulatory Economics, 43: 168-195.
Guerrero de Lizardi, Carlos (2006): “Una aproximación al sesgo de medición del precio de las
computadoras personales en México”, Economía Mexicana.
Hazlett, T. and C. Bazelon (2005), “Regulated Unbundling of Telecommunications Networks: A
Spetting Stone to Facilities-Based Competition”, mimeo.
Hoernig, S. (2001), “Collusion and confusion?”, mimeo.
Höffler, F. (2007), “Cost and benefits from infrastructure competition. Estimating welfare effects
from broadband access competition”, Telecommunications Policy, 31(6-7), 401-418.
Houpis, G., J. Lucena and J. Santamaría (2011) “Geographic Segmentation of Broadband Markets:
Appropiate Differentiation or Risk to a Single EU Market?”, Communications & Strategies, 82,
105-.
Laffont, J. and J. Tirole (2000), “Competition in Telecommunications”, Cambridge, Mass, The MIT Press.
Lee, S., M. Marcu and S. Lee (2011), “An Empirical analysis of fixed and mobile broadband
diffusion”, Information Economics and Policy, 23, 227-233.
Miravete, E. and L. Röller (2004), “Estimating markups under nonlinear pricing competition”,
Journal of the European Economic Association, 2, 526-536.
Nardotto, M., T. Valletti. and F. Verboven. (2012), “Unbundling the incumbent: Evidence from
UK broadband”, mimeo
Pereira, P. and T. and Ribeiro (2010), “The impact on broadband access to the Internet of the dual
ownership of telephone and cable networks”, International Journal of Industrial Organization,29, 283–
293.
Renda, Andrea, “Competition-regulation interface in telecommunications: What’s left of the
essential facility doctrine”. Telecommunications Policy, 34, 23-35.
Rosen, S (1974), “Hedonic Prices and Implicit Markets: Product Differentiation in Pure
Competition”, Journal of Political Economy, 34-55.
23
RTR (Austrian Telecoms Regulator).“Further Information about fixed-mobile Broadband
substitution in Austria (2010)”. RTR Working Paper.
Savage, S. and D. Waldman (2004), “United States Demand for Internet Access”. Review of Network
Economics, Vol.3, Issue 3 – September.
Schwarz, A. (2007), “Wholesale market definition in telecommunications: The issue of wholesale
broadband access”. Telecommunications Policy, 31, 251-264.
Vareda, J. and S. Hoernig (2007), The Race for Telecoms Infrastructures Investment with Bypass:
Can Access Regulation Achive The first best? CEPR Discussion Papers 6203.
Vogelsang, I. (2003), “Price Regulation of access to telecommunications networks”, Journal of
Economic Literature, 41 (3).
Wallsten, S. and J.L. Riso (2008), ”Residential and Business Broadband Prices: Part 1: An Empirical
Analysis of Metering and Other Price Determinants”, mimeo.
Waverman, L., M. Leschi, B. Reillier and K. Dasgupta (2007), “Access regulation and Infrastructure
Investment in the Telecommunications Sector: An empirical Investigation”, mimeo.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 26
Research Institute of Applied Economics Working Paper 2013/09 pag. 26
26
Llista Document de Treball
List Working Paper
WP 2013/09 “Broadband prices in the European Union: competition and commercial strategies” Calzada,
J. and Martínez, F.
WP 2013/08 “Double Penalty in Returns to Education: Informality and Educational Mismatch in the
Colombian Labour market” Herrera-Idárraga, P.; López-Bazo, E. and Motellón, E.
WP 2013/07 “Firm exports, innovation, … and regions” López-Bazo, E. and Motellón, E.
WP 2013/06 “Fiscal sustainability and fiscal shocks in a dollarized and oil-exporting country: Ecuador”
Marí Del Cristo, L.M. and Gómez-Puig, M.
WP 2013/05 “Informality and Overeducation in the Labor Market of a Developing Country” Herrera-
Idárraga, P.; López-Bazo, E. and Motellón, E.
WP 2013/04 “GLS based unit root tests for bounded processes” Carrion-i-Silvestre, J.Ll. and Gadea,
M.D.
WP 2013/03 “Determinants of Broadband Access: Is Platform Competition always the Key Variable to
Success?” Fageda, X.; Rubio, R. and Termes, M.
WP 2013/02 “Beyond Value-at-Risk: GlueVaR Distortion Risk Measures” Belles-Sampera, J.; Guillén, M.
and Santolino, M.
WP 2013/01 “How systemic is Spain for Europe?” Claeys, P. and Vašícek, B.
WP 2012/24 “Changes in Wage Structure in Mexico Going Beyond the Mean: An Analysis of Differences
in Distribution, 1987-2008” Tello, C.; Ramos, R, and Artís, M.
WP 2012/23 “Decomposing the Rural-Urban Differential in Student Achievement in Colombia Using
PISA Microdata” Ramos, R.; Duque, J.C. and Nieto, S.
WP 2012/22 “Recovery Risk and Labor Costs in Public-Private-Partnerships: Contractual Choice in the
U.S. Water industry” Albalate, D, Bel, G. and Geddes, R.
WP 2012/21 “Beyond pure public and pure private management models: Mixed firms in the European
Airport Industry” Albalate, D.; Bel, G. and Fageda, X.
WP 2012/20 “The determinants of contractual choice for private involvement in infrastructure projects in
the United States” Albalate, D, Bel, G. and Geddes, R.
WP 2012/19 “Measuring Sovereign Bond Spillover in Europe and the Impact of Rating News” Claeys, P.
and Vašíþek, B.
WP 2012/18 “Entry Regulation Asymmetries and Gasoline Competition in a Mixed Motorway Network”
Albalate, D. and Perdiguero, J.
WP 2012/17 “Regulation of Port Charges in Spain: Global versus Local Competition” Fageda, X. and
Gonzalez-Aregall, M.
WP 2012/16 “Pass-through in dollarized countries: should Ecuador abandon the U.S. Dollar?” Marí del
Cristo, M.L. and Gómez-Puig, M.
WP 2012/15 “A Note on the Relationship Between the Cyclicality of Markups and Fiscal Policy” Claeys,
P. and Costa, L.F.
WP 2012/14 “Do intra- and inter-industry spillovers matter? CDM model estimates for Spain” Goya, E.;
Vayá, E and Suriñach, J.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 27
Research Institute of Applied Economics Working Paper 2013/09 pag. 27
27
WP 2012/13 “What Drives the Choice of Partners in R&D Cooperation? Heterogeneity across Sectors”
Badillo, E. and Moreno, R.
WP 2012/12 “Corruption and local politics: does it pay to be a crook?” Jiménez, J.L. and García, C.
WP 2012/11 “Not always sunny in paradise: prices and brand diversity in touristic areas supermarkets”
Campos, J.; Jiménez, J.L. and Suárez-Alemán, A.
WP 2012/10 “The institutional, economic and social determinants of local government transparency”
Albalate, D.
WP 2012/09 “The business excellence attraction composite index (BEACI). Design and apllication to the
municipalities of the Barcelona province” Murillo, J.; Romaní, J.; Suriñach, J.
WP 2012/08 “Policy options for the promotion of electric vehicles: a review” Perdiguero, J. and Jiménez,
J.L.
WP 2012/07 “Price differences between domestic and international air markets: an empirical application
to routes from Gran Canaria” Fageda, X.; Jiménez, J.L. and Díaz Santamaría, C.
WP 2012/06 “Building a “quality in work” index in Spain” López-Tamayo, J.; Royuela, V. and Suriñach, J.
WP 2012/05 “Mergers and difference-in-difference estimator: why firms do not increase prices?”
Jiménez, J.L. and Perdiguero, J.
WP 2012/04 “What attracts knowledge workers? The role of space, social connections, institutions, jobs
and amenities” Miguélez, E. and Moreno, R.
WP 2012/03 “What Drives the Urban Wage Premium? Evidence along the Wage Distribution” Matano, A.
and Naticchioni, P.
WP 2012/02 “Location Patterns of Creative Capital and Regional Disparities in Spain” Kerimoglu, E. and
Karahasan, B.C.
WP 2012/01 “The connection between distortion risk measures and ordered weighted averaging
operators” Belles-Sampera, J.; Merigó, J.M.; Guillén, M. and Santolino, M.
WP 2011/26 “Productivity and innovation spillovers: Micro evidence from Spain” Goya, E.; Vayá, E. and
Suriñach, J.
WP 2011/25 “The regional distribution of unemployment. What do micro-data tell us?” López-Bazo, E.
and Motellón, E.
WP 2011/24 “Vertical relations and local competition: an empirical approach” Perdiguero, J.
WP 2011/23 “Air services on thin routes: Regional versus low-cost airlines” Fageda, X. and Flores-Fillol,
R.
WP 2011/22 “Measuring early childhood health: a composite index comparing Colombian departments”
Osorio, A.M.; Bolancé, C. and Alcañiz, M.
WP 2011/21 “A relational approach to the geography of innovation: a typology of regions” Moreno, R.
and Miguélez, E.
WP 2011/20 “Does Rigidity of Prices Hide Collusion?” Jiménez, J.L and Perdiguero, J.
WP 2011/19 “Factors affecting hospital admission and recovery stay duration of in-patient motor victims
in Spain” Santolino, M.; Bolancé, C. and Alcañiz, M.
WP 2011/18 “Why do municipalities cooperate to provide local public services? An empirical analysis”
Bel, G.; Fageda, X. and Mur, M.
WP 2011/17 “The "farthest" need the best. Human capital composition and development-specific
economic growth” Manca, F.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 28
Research Institute of Applied Economics Working Paper 2013/09 pag. 28
28
WP 2011/16 “Causality and contagion in peripheral EMU public debt markets: a dynamic approach”
Gómez-Puig, M. and Sosvilla-Rivero, S.
WP 2011/15 “The influence of decision-maker effort and case complexity on appealed rulings subject to
multi-categorical selection” Santolino, M. and Söderberg, M.
WP 2011/14 “Agglomeration, Inequality and Economic Growth” Castells, D. and Royuela, V.
WP 2011/13 “A correlation sensitivity analysis of non-life underwriting risk in solvency capital
requirement estimation” Bermúdez, L.; Ferri, A. and Guillén, M.
WP 2011/12 “Assessing agglomeration economies in a spatial framework with endogenous regressors”
Artis, M.J.; Miguélez, E. and Moreno, R.
WP 2011/11 “Privatization, cooperation and costs of solid waste services in small towns” Bel, G; Fageda,
X. and Mur, M.
WP 2011/10 “Privatization and PPPS in transportation infrastructure: Network effects of increasing user
fees” Albalate, D. and Bel, G.
WP 2011/09 “Debating as a classroom tool for adapting learning outcomes to the European higher
education area” Jiménez, J.L.; Perdiguero, J. and Suárez, A.
WP 2011/08 “Influence of the claimant’s behavioural features on motor compensation outcomes” Ayuso,
M; Bermúdez L. and Santolino, M.
WP 2011/07 “Geography of talent and regional differences in Spain” Karahasan, B.C. and Kerimoglu E.
WP 2011/06 “How Important to a City Are Tourists and Daytrippers? The Economic Impact of Tourism
on The City of Barcelona” Murillo, J; Vayá, E; Romaní, J. and Suriñach, J.
WP 2011/05 “Singling out individual inventors from patent data” Miguélez,E. and Gómez-Miguélez, I.
WP 2011/04 “¿La sobreeducación de los padres afecta al rendimiento académico de sus hijos?” Nieto,
S; Ramos, R.
WP 2011/03 “The Transatlantic Productivity Gap: Is R&D the Main Culprit?” Ortega-Argilés, R.; Piva, M.;
and Vivarelli, M.
WP 2011/02 “The Spatial Distribution of Human Capital: Can It Really Be Explained by Regional
Differences in Market Access?” Karahasan, B.C. and López-Bazo, E
WP 2011/01 “I If you want me to stay, pay” . Claeys, P and Martire, F
WP 2010/16 “Infrastructure and nation building: The regulation and financing of network transportation
infrastructures in Spain (1720-2010)”Bel,G
WP 2010/15 “Fiscal policy and economic stability: does PIGS stand for Procyclicality In Government
Spending?” Maravalle, A ; Claeys, P.
WP 2010/14 “Economic and social convergence in Colombia” Royuela, V; Adolfo García, G.
WP 2010/13 “Symmetric or asymmetric gasoline prices? A meta-analysis approach” Perdiguero, J.
WP 2010/12 “Ownership, Incentives and Hospitals” Fageda,X and Fiz, E.
WP 2010/11 “Prediction of the economic cost of individual long-term care in the Spanish population”
Bolancé, C; Alemany, R ; and Guillén M
WP 2010/10 “On the Dynamics of Exports and FDI: The Spanish Internationalization Process” Martínez-
Martín J.
WP 2010/09 “Urban transport governance reform in Barcelona” Albalate, D ; Bel, G and Calzada, J.
WP 2010/08 “Cómo (no) adaptar una asignatura al EEES: Lecciones desde la experiencia comparada
en España” Florido C. ; Jiménez JL. and Perdiguero J.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 29
Research Institute of Applied Economics Working Paper 2013/09 pag. 29
29
WP 2010/07 “Price rivalry in airline markets: A study of a successful strategy of a network carrier against
a low-cost carrier” Fageda, X ; Jiménez J.L. ; Perdiguero , J.
WP 2010/06 “La reforma de la contratación en el mercado de trabajo: entre la flexibilidad y la seguridad”
Royuela V. and Manuel Sanchis M.
WP 2010/05 “Discrete distributions when modeling the disability severity score of motor victims”
Boucher, J and Santolino, M
WP 2010/04 “Does privatization spur regulation? Evidence from the regulatory reform of European
airports . Bel, G. and Fageda, X.”
WP 2010/03 “High-Speed Rail: Lessons for Policy Makers from Experiences Abroad”. Albalate, D ; and
Bel, G.”
WP 2010/02 “Speed limit laws in America: Economics, politics and geography”. Albalate, D ; and Bel, G.”
WP 2010/01 “Research Networks and Inventors’ Mobility as Drivers of Innovation: Evidence from
Europe” Miguélez, E. ; Moreno, R. ”
WP 2009/26 ”Social Preferences and Transport Policy: The case of US speed limits” Albalate, D.
WP 2009/25 ”Human Capital Spillovers Productivity and Regional Convergence in Spain” , Ramos, R ;
Artis, M.; Suriñach, J.
WP 2009/24 “Human Capital and Regional Wage Gaps” ,López-Bazo,E. Motellón E.
WP 2009/23 “Is Private Production of Public Services Cheaper than Public Production? A meta-
regression analysis of solid waste and water services” Bel, G.; Fageda, X.; Warner. M.E.
WP 2009/22 “Institutional Determinants of Military Spending” Bel, G., Elias-Moreno, F.
WP 2009/21 “Fiscal Regime Shifts in Portugal” Afonso, A., Claeys, P., Sousa, R.M.
WP 2009/20 “Health care utilization among immigrants and native-born populations in 11 European
countries. Results from the Survey of Health, Ageing and Retirement in Europe” Solé-Auró,
A., Guillén, M., Crimmins, E.M.
WP 2009/19 “La efectividad de las políticas activas de mercado de trabajo para luchar contra el paro. La
experiencia de Cataluña” Ramos, R., Suriñach, J., Artís, M.
WP 2009/18 “Is the Wage Curve Formal or Informal? Evidence for Colombia” Ramos, R., Duque, J.C.,
Suriñach, J.
WP 2009/17 “General Equilibrium Long-Run Determinants for Spanish FDI: A Spatial Panel Data
Approach” Martínez-Martín, J.
WP 2009/16 “Scientists on the move: tracing scientists’ mobility and its spatial distribution” Miguélez, E.;
Moreno, R.; Suriñach, J.
WP 2009/15 “The First Privatization Policy in a Democracy: Selling State-Owned Enterprises in 1948-
1950 Puerto Rico” Bel, G.
WP 2009/14 “Appropriate IPRs, Human Capital Composition and Economic Growth” Manca, F.
WP 2009/13 “Human Capital Composition and Economic Growth at a Regional Level” Manca, F.
WP 2009/12 “Technology Catching-up and the Role of Institutions” Manca, F.
WP 2009/11 “A missing spatial link in institutional quality” Claeys, P.; Manca, F.
WP 2009/10 “Tourism and Exports as a means of Growth” Cortés-Jiménez, I.; Pulina, M.; Riera i
Prunera, C.; Artís, M.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 30
Research Institute of Applied Economics Working Paper 2013/09 pag. 30
30
WP 2009/09 “Evidence on the role of ownership structure on firms' innovative performance” Ortega-
Argilés, R.; Moreno, R.
WP 2009/08 “¿Por qué se privatizan servicios en los municipios (pequeños)? Evidencia empírica sobre
residuos sólidos y agua” Bel, G.; Fageda, X.; Mur, M.
WP 2009/07 “Empirical analysis of solid management waste costs: Some evidence from Galicia, Spain”
Bel, G.; Fageda, X.
WP 2009/06 “Intercontinental fligths from European Airports: Towards hub concentration or not?” Bel, G.;
Fageda, X.
WP 2009/05 “Factors explaining urban transport systems in large European cities: A cross-sectional
approach” Albalate, D.; Bel, G.
WP 2009/04 “Regional economic growth and human capital: the role of overeducation” Ramos, R.;
Suriñach, J.; Artís, M.
WP 2009/03 “Regional heterogeneity in wage distributions. Evidence from Spain” Motellón, E.; López-
Bazo, E.; El-Attar, M.
WP 2009/02 “Modelling the disability severity score in motor insurance claims: an application to the
Spanish case” Santolino, M.; Boucher, J.P.
WP 2009/01 “Quality in work and aggregate productivity” Royuela, V.; Suriñach, J.
WP 2008/16 “Intermunicipal cooperation and privatization of solid waste services among small
municipalities in Spain” Bel, G.; Mur, M.
WP 2008/15 “Similar problems, different solutions: Comparing refuse collection in the Netherlands and
Spain” Bel, G.; Dijkgraaf, E.; Fageda, X.; Gradus, R.
WP 2008/14 “Determinants of the decision to appeal against motor bodily injury settlements awarded by
Spanish trial courts” Santolino, M
WP 2008/13 “Does social capital reinforce technological inputs in the creation of knowledge? Evidence
from the Spanish regions” Miguélez, E.; Moreno, R.; Artís, M.
WP 2008/12 “Testing the FTPL across government tiers” Claeys, P.; Ramos, R.; Suriñach, J.
WP 2008/11 “Internet Banking in Europe: a comparative analysis” Arnaboldi, F.; Claeys, P.
WP 2008/10 “Fiscal policy and interest rates: the role of financial and economic integration” Claeys, P.;
Moreno, R.; Suriñach, J.
WP 2008/09 “Health of Immigrants in European countries” Solé-Auró, A.; M.Crimmins, E.
WP 2008/08 “The Role of Firm Size in Training Provision Decisions: evidence from Spain” Castany, L.
WP 2008/07 “Forecasting the maximum compensation offer in the automobile BI claims negotiation
process” Ayuso, M.; Santolino, M.
WP 2008/06 “Prediction of individual automobile RBNS claim reserves in the context of Solvency II”
Ayuso, M.; Santolino, M.
WP 2008/05 “Panel Data Stochastic Convergence Analysis of the Mexican Regions” Carrion-i-Silvestre,
J.L.; German-Soto, V.
WP 2008/04 “Local privatization, intermunicipal cooperation, transaction costs and political interests:
Evidence from Spain” Bel, G.; Fageda, X.
WP 2008/03 “Choosing hybrid organizations for local services delivery: An empirical analysis of partial
privatization” Bel, G.; Fageda, X.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 31
Research Institute of Applied Economics Working Paper 2013/09 pag. 31
31
WP 2008/02 “Motorways, tolls and road safety. Evidence from European Panel Data” Albalate, D.; Bel,
G.
WP 2008/01 “Shaping urban traffic patterns through congestion charging: What factors drive success or
failure?” Albalate, D.; Bel, G.
WP 2007/19 “La distribución regional de la temporalidad en España. Análisis de sus determinantes”
Motellón, E.
WP 2007/18 “Regional returns to physical capital: are they conditioned by educational attainment?”
López-Bazo, E.; Moreno, R.
WP 2007/17 “Does human capital stimulate investment in physical capital? evidence from a cost system
framework” López-Bazo, E.; Moreno, R.
WP 2007/16 “Do innovation and human capital explain the productivity gap between small and large
firms?” Castany, L.; López-Bazo, E.; Moreno, R.
WP 2007/15 “Estimating the effects of fiscal policy under the budget constraint” Claeys, P.
WP 2007/14 “Fiscal sustainability across government tiers: an assessment of soft budget constraints”
Claeys, P.; Ramos, R.; Suriñach, J.
WP 2007/13 “The institutional vs. the academic definition of the quality of work life. What is the focus of
the European Commission?” Royuela, V.; López-Tamayo, J.; Suriñach, J.
WP 2007/12 “Cambios en la distribución salarial en españa, 1995-2002. Efectos a través del tipo de
contrato” Motellón, E.; López-Bazo, E.; El-Attar, M.
WP 2007/11 “EU-15 sovereign governments’ cost of borrowing after seven years of monetary union”
Gómez-Puig, M..
WP 2007/10 “Another Look at the Null of Stationary Real Exchange Rates: Panel Data with Structural
Breaks and Cross-section Dependence” Syed A. Basher; Carrion-i-Silvestre, J.L.
WP 2007/09 “Multicointegration, polynomial cointegration and I(2) cointegration with structural breaks.
An application to the sustainability of the US external deficit” Berenguer-Rico, V.; Carrion-i-
Silvestre, J.L.
WP 2007/08 “Has concentration evolved similarly in manufacturing and services? A sensitivity analysis”
Ruiz-Valenzuela, J.; Moreno-Serrano, R.; Vaya-Valcarce, E.
WP 2007/07 “Defining housing market areas using commuting and migration algorithms. Catalonia
(Spain) as an applied case study” Royuela, C.; Vargas, M.
WP 2007/06 “Regulating Concessions of Toll Motorways, An Empirical Study on Fixed vs. Variable Term
Contracts” Albalate, D.; Bel, G.
WP 2007/05 “Decomposing differences in total factor productivity across firm size” Castany, L.; Lopez-
Bazo, E.; Moreno, R.
WP 2007/04 “Privatization and Regulation of Toll Motorways in Europe” Albalate, D.; Bel, G.; Fageda, X.
WP 2007/03 “Is the influence of quality of life on urban growth non-stationary in space? A case study of
Barcelona” Royuela, V.; Moreno, R.; Vayá, E.
WP 2007/02 “Sustainability of EU fiscal policies. A panel test” Claeys, P.
WP 2007/01 “Research networks and scientific production in Economics: The recent spanish
experience” Duque, J.C.; Ramos, R.; Royuela, V.
WP 2006/10 “Term structure of interest rate. European financial integration” Fontanals-Albiol, H.; Ruiz-
Dotras, E.; Bolancé-Losilla, C.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 32
Research Institute of Applied Economics Working Paper 2013/09 pag. 32
32
WP 2006/09 “Patrones de publicación internacional (ssci) de los autores afiliados a universidades
españolas, en el ámbito económico-empresarial (1994-2004)” Suriñach, J.; Duque, J.C.;
Royuela, V.
WP 2006/08 “Supervised regionalization methods: A survey” Duque, J.C.; Ramos, R.; Suriñach, J.
WP 2006/07 “Against the mainstream: nazi privatization in 1930s germany” Bel, G.
WP 2006/06 “Economía Urbana y Calidad de Vida. Una revisión del estado del conocimiento en
España” Royuela, V.; Lambiri, D.; Biagi, B.
WP 2006/05 “Calculation of the variance in surveys of the economic climate” Alcañiz, M.; Costa, A.;
Guillén, M.; Luna, C.; Rovira, C.
WP 2006/04 “Time-varying effects when analysing customer lifetime duration: application to the
insurance market” Guillen, M.; Nielsen, J.P.; Scheike, T.; Perez-Marin, A.M.
WP 2006/03 “Lowering blood alcohol content levels to save lives the european experience” Albalate, D.
WP 2006/02 “An analysis of the determinants in economics and business publications by spanish
universities between 1994 and 2004” Ramos, R.; Royuela, V.; Suriñach, J.
WP 2006/01 “Job losses, outsourcing and relocation: empirical evidence using microdata” Artís, M.;
Ramos, R.; Suriñach, J.
Institut de Recerca en Economia Aplicada Regional i Pública Document de Treball 2013/09 pàg. 33
Research Institute of Applied Economics Working Paper 2013/09 pag. 33
33
doc_524182775.pdf