Data Governance in Connected Cars

AuthorWolfgang Kerber
PositionProfessor of Economics, Marburg Centre of Institutional Economics, School of Business & Economics, University of Marburg, Germany
Pages310-331
2018
Wolfgang Kerber
310
3
Data Governance in Connected Cars
The Problem of Access to In-Vehicle Data
by Wolfgang Kerber*
© 2018 Wolfgang Kerber
Everybody may disseminate this ar ticle by electronic m eans and make it available for downloa d under the terms and
conditions of the Digital P eer Publishing Licence (DPPL). A copy of the license text may be obtain ed at http://nbn-resolving.
de/urn:nbn:de:0009-dppl-v3-en8.
Recommended citation: Wolf gang Kerber, Data Governance in Connected Cars: The Problem of A ccess to In-Vehicle Data, 9
(2018) JIPITEC 310 para 1.
Keywords: Data governance; connected cars; data economy; data access
the ecosystem of connected and automated mobility.
The paper offers an overview about this policy dis-
cussion and analyzes this problem from an economic
perspective by utilizing a market failure analysis. Be-
sides competition problems (especially on markets
for aftermarket and other services in the connected
car) and market failures in regard to technological
choice (extended vehicle vs. interoperable on-board
application platform), information and privacy prob-
lems (“notice and consent” solutions) can emerge,
leading to the question of appropriate regulatory so-
lutions. The paper discusses solutions through data
portability, data rights, competition law, and recom-
mends a sector-specific regulatory approach.
Abstract: Through the application of the tech-
nological solution of the “extended vehicle” concept,
the car manufacturers can capture exclusive con-
trol of the data of connected cars leading to serious
concerns about negative effects on competition, in-
novation and consumer choice on the markets for
aftermarket and other complementary services in
the ecosystem of connected and automated driv-
ing. Therefore, a controversial policy discussion has
emerged in the EU about access to in-vehicle data
and the connected car for independent service pro-
viders in the automotive industry. This paper claims
that this problem should be seen as part of the gen-
eral question of the optimal governance of data in
A. Introduction
1 Connected, automated (and later autonomous) cars
can lead to large benets both to users of cars and
to society, such as more convenience, reduction of
accidents, congestion and emissions. Connected and
automated driving is a technological revolution not
only for the automotive industry (and their business
models) but also for the mobility in society. Therefore,
a policy discussion has emerged in the EU and within
the Member States on how to enable connected and
automated driving. The recent EU Communication
“On the road to automated mobility: An EU strategy
for mobility of the future” offers a broad overview
about the challenges and problems that have to be
solved.1 There are many open regulatory questions
* Professor of Economics, Marburg Centre of Institutional
regarding safety and cybersecurity risks, liability
problems, ethical questions, standardization and
interoperability problems, privacy concerns, and
the governance of data, especially data access.
2
This article focuses on the question of the governance
of the huge mass of data produced in connected cars.
An important part of this data governance problem
is the current controversial policy discussion
Economics, School of Business & Economics, University of
Marburg, Germany; Email: kerber@wiwi.uni-marburg.de.
1 See EU Commission, A European strategy on Cooperative
Intelligent Transport Systems, a milestone towards
cooperative, connected and automated mobility, 30.11.2016,
COM(2016) 766 n.; EU Commission, On the road to
automated mobility: An EU strategy for mobility of the
future, 17.5.2018, COM(2018) 283 n.; Bundesregierung,
Strategy for Automated and Connected Driving, 2015.
Data Governance in Connected Cars
2018
311
3
about “access to in-vehicle data and resources”
for independent providers of services within the
ecosystem of connected and automated mobility.2
The car manufacturers (OEMs: original equipment
manufacturers) use the so called “extended vehicle
concept” that implies transmitting all data produced
in the car directly to proprietary servers of the OEMs
granting them an exclusive (“monopolistic”) control
of these data. Many rms within the ecosystem of
connected and automated mobility could provide
a wide range of services to the cars owners and
drivers if they also have access to the in-vehicle
data. These independent service providers – as well
as consumer associations – are concerned that this
“privileged” position of the OEMs allows them to
control the automotive aftermarkets and adjacent
services leading to less competition, less consumer
choice and less innovation. Therefore, the current
policy discussion focuses on this conict between
the OEMs, who defend their extended vehicle
concept with safety and security arguments, and
the many independent service providers, who
demand regulatory solutions regarding access to in-
vehicle data and connected cars for ensuring fair and
undistorted competition concerning the provision of
services in the ecosystem of connected driving. The
most important proposals are either – in the short-
term - a non-discriminatory governance solution for
the in-vehicle data (e.g., a “shared server”) or in the
long-term, the transition to another technological
solution (on-board application platform), which
would allow the car owners to control access
to in-vehicle data and the car. Although the EU
Commission acknowledges the problem that the
“centralisation of in-vehicle” data in the extended
vehicle concept might trigger a competition problem
and wants to improve access to these data, so far
only a recommendation with guidance on non-
binding principles for access to in-vehicle data has
been planned.3
3 Although the current policy discussion is primarily
about access to in-vehicle data and resources for
independent service providers, the problem of
nding an appropriate governance solution for
data in the ecosystem of connected and automated
mobility is a much more complex problem. One
important problem is the fact that most in-vehicle
data are also personal data that are subject to the
requirements of EU data protection law. Due to non-
rivalry in the use of data - i.e. that many rms can
use the same data for their services and innovations
2 See C-ITS Platform, Final Report, 2016; TRL, Access to In-
Vehicle Data and Resources – Final Report, 18.05.2017;
and as overview Specht/Kerber, Datenrechte – Eine
rechts- und sozialwissenschaftliche Analyse im Vergleich
Deutschland – USA, 2018, available at:
de/de/blog-item/gutachten-datenrechte-eine-rechts-und-
sozialwissenschaftliche-analyse-im-vergleich>, 169-192.
3 EU Commission 2018 (n 1) 13.
- the question arises whether an exclusive control
of in-vehicle data through one stakeholder in such
a complex ecosystem of connected driving with
so many different stakeholders is an economically
efcient governance solution for these data. Or
should a more sophisticated data governance
solution be chosen, which allows more stakeholders
to get access to these data as a valuable input for
their services and innovations? This economics of
data perspective is directly linked to the recent
discussion about data rights and the efforts of the
EU Commission for better data access and reuse.
4
However, any solution has to also comply with EU
data protection law for protecting the privacy of
the car users. This article claims that the problem
of access to in-vehicle data should be seen as part
of the more general question concerning how a
comprehensive governance solution for the data
that are produced in the ecosystem of connected
and automated mobility should look like.
4
The objective of this article is to provide (1) an
overview about the current discussion about access
to data in the connected car (section B), (2) an
economic analysis of the data governance problem
that asks for potential market failure problems
(section C), and (3) a discussion about possible policy
approaches for dealing with the data governance
problems (section D).
5
The analytical approach used in this article is an
economic analysis of potential market failures that
can arise in the ecosystem of connected driving and
which might make regulatory activities necessary
for solving the data governance problems. One of the
potential market failure problems are certainly the
competition problems that might be caused by the
exclusive control of in-vehicle data in the extended
vehicle concept on the markets for aftermarket and
complementary services. In that respect, an analysis
of competition between OEMs is also necessary.
A second potential market failure refers to the
question of whether it can be expected that OEMs
choose technological solutions that are optimal for
the entire ecosystem of connected and automated
driving, such as, the extended vehicle concept or
the on-board application platform. Based upon the
insights of the economics of interoperability and
standardization, serious doubts arise concerning
whether OEMs have the right incentives for making
optimal technological decisions. An additional third
concern is that car users as consumers might run
4 See EU Commission, Building a European data economy,
10.1.2017, COM(2017) 9 n.; EU Commission, Towards
a common European data space, 25.4.2018, COM(2018)
232 n.; and as overview Kerber, Rights on Data: The EU
Communication “Building a European Data Economy”
from an Economic Perspective, in: Lohsse/Schulze/
Staudenmayer, Trading Data in the Digital Economy: Legal
Concepts and Tools, 2017, 109-133.
2018
Wolfgang Kerber
312
3
into similar problems of protecting and dealing
properly with their personal data and their privacy
as they are well-known with respect to other
internet service providers, where it is doubtful
whether and to what extent consumers can make
well-informed rational decisions about the provision
of data to digital companies. In all three cases the
preliminary assessment in this paper suggests that
serious market failures can exist, although much
more research is necessary. Therefore, the results of
this analysis raise serious doubts about the currently
used extended vehicle concept of the OEMs, which
might be both a wrong technological solution,
especially in the long term, and lead to negative
effects regarding competition on markets for
aftermarket and complementary services. It will also
be shown that safety and security concerns cannot
justify the exclusive control of data of OEMs and their
power to appropriate the value of in-vehicle data
through this monopolistic gatekeeper position. The
development of an on-board application platform (as
an open interoperable telematics platform) would
avoid many of the disadvantages of the extended
vehicle concept and might also be more compatible
with the needs of the long-term architecture of an
integrated ecosystem of connected and automated
mobility.
6 Due to the complexity of the technological and data
governance problem, this article cannot offer a
clear-cut policy proposal with regards to connected
driving. However, in an overview about recent
discussions of possible policy approaches to solve
data access and data governance solutions, section D
discusses the right to data portability (Art. 20 GDPR),
the general introduction of explicit data rights in
civil law, as well as possible solutions in competition
law, for example, data access rights as remedies
for the refusal to grant access to data as abusive
behavior of rms with market power (as, e.g., Art.
102 TFEU). However, this article concludes with the
suggestion that due to the large complexity of this
problem, looking for a sector-specic regulatory
solution might be the most suitable path for solving
the data governance problem in the ecosystem of
connected and automated driving.
B. Access to in-vehicle data
and resources: A policy
discussion in the EU
7 In the connected and automated car many different
kinds of data are produced, particularly through
sensors. This can be technical data regarding the
car and its components, data about the road, weather
and trafc conditions, the driving behavior of the car
drivers, location data, as well as data concerning the
use of entertainment, navigation and many other
services by the car users. Through the connectivity
of the car via mobile communication, these data
can be transmitted in real-time to external entities,
for example, to an external server of the OEMs,
but also a direct exchange of data is possible that
allows the downloading of software and updates. The
connectivity and the in-vehicle data allow for many
new (and innovative) services that can be offered to
car users. They can include new forms of repair and
maintenance services such as remote diagnostics
and maintenance, navigation services, parking
apps, search services for hotels and restaurants,
entertainment, online-shopping, as well as new
insurance schemes (used-based insurance), among
others.5 The providers of these services however
often need access to the in-vehicle data and/or
to the connected car for providing these services
(and for communication with the car users) for
being capable to enter the markets for aftermarket
and complementary services.
6
A part of these new
services would also require real-time access to these
data and the car.7
8
As part of its “Cooperative Intelligent Transport
Systems” initiative for solving problems of connected
and automated driving, the EU Commission has
brought together all stakeholders on the C-ITS
5 See generally about connected and automated cars OECD/
ITF, Automated and Autonomous Driving. Regulation under
uncertainty. Corporate Partnership Report, 2015; Anderson
et al., Autonomous Vehicle Technology – A Guide for
Policymakers, 2016; Alonso Raposo et al., The revolution of
driving: from Connected Vehicles to Coordinated Automated
Road Transport (C-ART), European Commission JRC Science
for Policy Report, Part I: Framework for a safe & efcient
Coordinated Automated Road Transport (C-ART) system,
2017; for the new business opportunities through the
connected car see McKinsey, Competing for the connected
customer: Perspectives on the opportunities created by
car connectivity and automation, McKinsey & Company,
Advanced Industries, September 2015; McKinsey, Car data:
Paving the way to value-creating mobility – Perspectives on
a new automotive business model, McKinsey & Company,
Advanced Industries, March 2016; BVDW, Connected Cars
– ein Diskussionspapier zum Thema Services, 2015; BVDW,
Connected Cars – Geschäftsmodelle. Diskussionspapier,
23.05.2016; BVDW, Connected Cars – Chancen und Risiken
für die künftigen Anbieter im Automobilmarkt, 2016.
6 The data of connected cars are also interesting for public
authorities, e.g. for trafc safety and regulation or law
enforcement.
7 Access to the connected car means independent service
providers have mobile access to (1) the IT system of the
car for either downloading data (“read”) or also uploading
data or providing services in the connected car (“write”)
as remote diagnosis or software updates, and (2) the
Human-Machine-Interface (HMI or dashboard) for direct
communication with the car drivers. If OEMs control this
access, they can block direct interaction between car drivers
and independent service providers. See for the technical
details TRL (n 2) 75-92; Martens/Mueller-Langer, Access
to digital car data and competition in aftersales services,
Digital Economy Working Paper 2018-0X, JRC Technical
Reports, 2018, 7-10.
Data Governance in Connected Cars
2018
313
3
platform.8 In this context the problem of access to
these data for independent service providers was
already discussed very clearly. An important result
for the ensuing policy discussion was a consensus
regarding ve guiding principles that should
apply to access to in-vehicle data. Besides solving
safety and security problems such as “tamper-
proof access and liability”, the compliance with
data protection and data privacy, and standardized
access / interoperability for facilitating use of the
same vehicle data, two other important principles
were introduced: The right of car users to decide
if data are provided and to whom (consent), and
that “all service providers should be in an equal,
fair, reasonable and non-discriminatory position to
offer services” to the car users - “fair and undistorted
competition”.9 Especially in the Working group 6 of
the C-ITS platform, which dealt with technological
solutions about access to in-vehicle data, the conict
between OEMs and independent service providers
became very apparent,10 because – as we will see
later in more detail (section C.II.) – technological
solutions can deeply inuence the governance of
data.
9 On the C-ITS platform three technological solutions
were discussed. For the following analysis and
discussion, it is sufcient to focus on two basic
technological solutions.
11
The rst one, the “external
server” solution, implies that all in-vehicle data are
transmitted to an external server (outside of the
car) and access to these data is only possible via this
external server. The “extended vehicle” concept of
the OEMs is one variant of this “external server”
solution, in which this is a proprietary server of the
OEMs that lead to their exclusive control of the data.12
Another variant of the “external server” solution is
the “shared server” concept. It is technologically the
same solution, however is not under the exclusive
control of the OEMs but under the governance of
a neutral entity that can give access to these data
to all stakeholders on non-discriminatory terms.
The second main technological solution is the “on-
board application platform”. In this solution the car
itself would be the platform on which the data are
stored, and the car owners can decide directly whom
8 See EU Commission 2016 (n 1); C-ITS Platform (n 2).
9 For these ve principles, see C-ITS Platform (n 2) 75.
10 See C-ITS Platform (n 2) 78-89.
11 For an explanation and analysis of the technological
solutions, see C-ITS Platform (n 2) 72-90; TRL (n 2) 32-49; and
Martens/Mueller-Langer (n 7), 7-13. The third solution, the
“in-vehicle interface”, is the currently existing On-Board
Diagnostic (OBD) Adapter, which is used for transmitting
data for emissions control and repair and maintenance
services. However, it is not such a basic solution as the two
solutions described in the following.
12 In recent publications this variant has also been called
“central data server platform” (Martens/Mueller-Langer [n
7] 8) or “centralization of in-vehicle data” (EU Commission
2018 [n 1] 13).
to grant access to the in-vehicle data and who can
get access to the car for providing services to the
car users. Since this technological solution leads to a
much more “open” version of the connected car, this
solution can also be seen as an open interoperable
telematic platform. Thus, the technological choice
between these two basic options is important (1) for
the question who has control of the data, and (2) for
the choice between a more interoperable “open” or
a more “closed” model of connected cars.13
10
In the following, the positions of the OEMs and
the independent service providers in this policy
discussion are briey summarized.14 The European
car manufacturers are mainly using the extended
vehicle concept in their connected cars and are
claiming via their associations that this model is
the only suitable model for access to in-vehicle data
and the connected car.15 The main argument of the
OEMs is that the exclusive control of the access to in-
vehicle data and the car is necessary, because it is the
only way to ensure the very high standard of safety
and security that is necessary for connected (and
automated) cars. Due to the risks of cyber-attacks,
manipulation, compromising the integrity of the
functions of the connected cars etc., all technological
solutions that would allow a direct exchange of data
with independent service providers would be too
dangerous for the safety and security of the car. The
responsibility of the OEMs for safety and security
is also directly linked to their liability regarding
the connected car. Therefore, safety and security
concerns are the reason why the connected car has
to be a closed system under the exclusive control of
the OEMs. The car manufacturers claim that they
are willing to grant access to in-vehicle data on their
proprietary servers, but only on the basis of freely
negotiated B2B-contracts with independent service
providers, and in that respect they will distinguish
different categories of data.
16
The OEMs also claim
that the extended vehicle concept allows the car
owner to freely choose between all service providers
that have contracts with the OEM with regards to
13 For a very detailed analysis of the advantages and problems
of an “open” vs. a “closed” model of connected cars see
Determann/Perens, Open Cars, Berkeley Technology Law
Journal, 2017, 915-988.
14 For a deeper analysis of the positions and arguments of
the different groups of stakeholders in the ecosystem of
connected driving, see Specht/Kerber (n 2) 49-55.
15 See, in particular, ACEA, Access to vehicle data for third-
party services. ACEA Position Paper, Brussels, December
2016a; ACEA, ACEA Strategy Paper on Connectivity, Brussels,
April 2016b; and VDA, Position. Zugang zum Fahrzeug und
zu im Fahrzeug generierten Daten. Berlin, 2016.
16 See VDA (n 15) 6 et seq. There are three exceptions
regarding access to data via free B2B-contracts: personal
data only with explicit consent of the car owners, repair and
maintenance information according to the regulated access
of the type approval regulation (see below in this section),
and anonymized data for the improvement of trafc safety
for public authorities.
2018
Wolfgang Kerber
314
3
necessary data or access to the car. It is less clear
how the OEMs defend their exclusive control of the
in-vehicle data. Only rather general remarks about
the huge investments into the development and the
operating costs of connected cars can be found in the
position papers.17
11
Despite the heterogeneity of the different groups
of independent service providers, there is a
large consensus concerning their critique of the
OEMs and their extended vehicle concept and
possible solutions.18 Repair and maintenance
service providers in particular, have emphasized
the importance of access to in-vehicle data and
the possibility to use their knowhow directly in
the vehicle, i.e. that they can get direct access to
the connected car.19 Especially important is that
independent service providers can develop and offer
many new innovative services (such as, e.g. remote
monitoring and maintenance) in the automotive
aftermarkets and in markets for complementary
services.20 They are concerned that the exclusive
control of OEMs regarding access to in-vehicle data
and the connected car can impede competition and
innovation on these markets. Access via an external
server can also impede innovation because certain
new services need access in real-time (whereas the
external server leads to time lags). Moreover, access
to raw data and not only aggregated or already
processed data can be important for new innovative
services. Another alleged problem is that OEMs can
observe the data access of independent service
providers and therefore monitor their transactions
with the car owners. These data can lead to a
competitive advantage of the OEMs concerning their
own services to the car users.
21
These arguments are
also relevant for many other independent service
17 See ACEA 2016b (n 16) 7. Very interesting but not clearly
elaborated are also hints about the danger of market
dominance through large tech companies if data are made
as accessible as possible according to the principle of “free
ow of data” (ibid, 1).
18 See FIGIEFA, Commission Communication on “Free
Flow of Data”. Input from the Independent Automotive
Aftermarket, 23 December 2016, AFCAR, Insurance, leasing,
dealers, vehicle inspection, automotive aftermarket and
consumers coalition: Keeping the principles of the Treaty
of Rome alive in the automotive digital age, Press Release,
Brussels, 23 March 2017, ADPA et al, EC Mobility Package
outlines vision for automated mobility but fails to set out a
clear plan for access to in-vehicle data. Press statement, 17
May 2018.
19 See for the following FIGIEFA (n 18) 14-17.
20 See FIGIEFA (n 18) 3: “Foreseeable use cases are for example
the proactive monitoring of safety-critical vehicle systems,
the predictive ... maintenance in the workshop, remote
monitoring of operations to prevent defects, remote
maintenance through software updates or reconguration
and automated services in case of a breakdown on the
road”.
21 See FIGIEFA (n 18) 14. These and other critical arguments
have already been discussed clearly in the Working Group 6
of the C-ITS platform (n 2, 78).
providers. Therefore, nearly all other stakeholders
reject the extended vehicle concept and the
“privileged” position of the OEMs, and demand in the
short term, non-discriminatory access to the data,
and, in the long term, the transition to an “open
telematics system“ (on-board application platform)
that would give the car users direct control of the
access to the data and the connected car. This is also
in line with the position of the consumer associations
who insist on fair and undistorted competition with
regards to aftermarket and complementary services
and the right of car users to choose freely between
all service providers (consumer choice).22 The
consumer associations also demand a clarication
about the rights of car owners with regards to the
data - including the non-personal data.23
12 Parallel to this policy discussion, the EU enacted a
reform of the type approval regulation for vehicles
in 2018.24 For a long time, competition law had to deal
with strategies of the OEMs that tried to foreclose
independent service providers from the often
highly protable automotive aftermarkets. Since the
1980s, the EU competition policy had implemented
regulatory provisions; rst in a sector-specic block
exemption regulation, and since 2007 in the type
approval regulation of vehicles that should ensure
that competition on these automotive aftermarkets
between the OEMs and the independent service
providers is not distorted or eliminated through a
lack of access to necessary technical information
for repair and maintenance services.25 The main
regulatory instrument for achieving this objective
was the introduction of an obligation of the
OEMs to grant the same access about necessary
22 See BEUC, Protecting European Consumers with connected
and automated cars. Position paper, Brussels, 11.12.2017;
FIA, Policy Position on Car Connectivity, Brussels, 2016a;
FIA, What Europeans think about connected cars, Brussels,
January 2016b.
23 See, e.g., BEUC (n 22) 8. In a survey of European car owners
about connected cars, 90% of the participants said that the
data produced in connected cars should be “owned” by the
car owners or the car drivers. See FIA 2016b (n 22) 1.
24 Regulation (EU) 2018/858 of the European Parliament and
of the Council of 30 May 2018 on the approval and market
surveillance of motor vehicles and their trailers, and of
systems, components and separate technical units intended
for such vehicles, amending Regulations (EC) No 715/2007
and (EC) No 595/2009 and repealing Directive 2007/46/EC.
25 Regulation (EC) 715/2007 of the European Parliament
and of the Council of 20 June 2007 on the type approval
of motor vehicles with respect to emissions from light
passenger and commercial vehicles (Euro 5 and Euro 6) and
on access to vehicle repair and maintenance information.
Regulation (EU) 461/2010 of 27 May 2010 on the application
of Article 101(3) of the Treaty on the Functioning of the
European Union to categories of vertical agreements and
concerted practices in the motor vehicle sector. See for
this regulation also Becker/Simon, GVO Nr. 461/2010
(Kfz-GVO) Vertriebs- und Kundendienstvereinbarungen
im Kfz-Sektor, in: Bornkamm/Montag/Säcker, Münchner
Kommentar Europäisches und Deutsches Wettbewerbsrecht
(Kartellrecht), 2015, 1173-1234.
Data Governance in Connected Cars
2018
315
3
information for repair and maintenance services
to independent service providers as to their own
service providers. Therefore, concerning repair
and maintenance services, independent providers
have already regulated access rights to essential
technical information and diagnostic data via the
on-board diagnostic (OBD) adapter for a long time.
This solution of “regulated access” to necessary
technical information for safeguarding fair and
undistorted competition on the aftermarkets for
repair and maintenance service providers, is a
broadly accepted regulation that has fullled its
task successfully.26 Although the reform of the type
approval regulation was triggered by the emissions
scandal in the automotive industry, it also led to
some adaptation of the rules about this regulated
access of independent service providers. This reform
did not take into account all the implications of the
transition from traditional to connected cars, but
the extent of the regulated access of independent
service providers regarding data and the car under
these new technological conditions, e.g. also for
providing new services (by using remote access),
was also discussed in this context. In the end, the
respective changes in the type approval regulation
have remained rather limited, but it is clear that this
regulatory access solution will be subject to further
regulatory discussions in the future with the same
conict between the OEMs and independent service
providers.27
13
In the C-ITS platform discussions, as the conict
between OEMs and independent service providers
about access to in-vehicle data and resources could
not be resolved between the stakeholders, it was a
logical next step that the EU Commission initiated
a study aiming to investigate to what extent the
different technological solutions are compatible
with the above-mentioned ve C-ITS guiding
principles about access to in-vehicle data and
resources (TRL 2017). This (the most comprehensive
to date) study about this access problem led to the
following results.
28
All technological solutions are
technically and legally feasible (also when it comes
to safety and security), but they each have different
advantages and problems. Although no solution
is superior, the study comes to the conclusion
that the “on-board application platform” is the
relatively best solution. Particularly important for
this result is that the extended vehicle concept is
assessed as being incompatible with the principle
of fair and undistorted competition. The study
26 See European Commission, Study on the operation of
the system of access to vehicle repair and maintenance
information, Final report, 2014.
27 Important changes of the type approval regulation refer to
rules about the support of repair and maintenance services
through wireless networks and the access to remote
diagnosis services of the OEMs.
28 See TRL (n 2) 8-16.
discusses a number of possible policy measures
which differ with regards to the time horizon and
the depth of policy intervention. This discussion
clearly suggests that in the short-term (under the
current technological “external server” solution),
the variant of the “shared server” would lead
to more compatibility with the principle of fair
and undistorted competition. However, in order
to ensure a far-reaching compatibility with this
principle, the interoperable on-board application
platform is recommended in the long-term.
The study acknowledges the safety and security
challenges of this solution but deems them to be
solvable. The study recommends encouraging the
development of a single interoperable platform,
but in the end does not go so far as to recommend
making such a platform mandatory for the OEMs.29
14 What is the state of the current policy discussion?
Despite the results of the TRL study, the conict
between OEMs and the independent service
providers could not be resolved. Whereas the
independent service providers still demand
legislative action, especially concerning a “shared
server” and interoperable platforms solutions, the
OEMs reject legislative measures and want to stick
to their extended vehicle concept. In February
2018, the European Parliament demanded that the
Commission publishes a legislative proposal on
access to in-vehicle data and resources with the
explicit objectives of maximum security and a level-
playing-eld for access for all third-parties “… to
protect consumer rights, promote innovation and
ensure fair, non-discriminatory competition on this
market …”.30 In its Communication “On the road to
automated mobility” (May 2018), the Commission
acknowledged the competition problems and
that the “centralisation of in-vehicle” data in the
extended vehicle concept might “not be sufcient
to ensure fair and undistorted competition between
service providers”.31 However, the Commission
seems to be reluctant to address this problem, and
therefore is not planning legislative actions with
binding rules. It rather wants to solve the problems
by publishing a recommendation with “guidance
on a data governance framework for access to and
sharing of data generated by connected vehicles”
based upon non-binding principles.32
29 See TRL (n 2) 160.
30 See EP, Report on a European strategy on Cooperative
Intelligent Transport Systems (2017/2067(INI)). Committee
on Transport and Tourism (PE610.712v02-00), 2018, 10.
31 EU Commission 2018 (n 1) 13.
32 See EU Commission, Roadmap Cooperative, Connected
and Automated Mobility (CCAM), Ref. Ares(2018)5386378 –
19/10/2018, 2.
2018
Wolfgang Kerber
316
3
C. Data Governance in Connected
Cars: An Economic Analysis of
Potential Market Failures
I. Introduction
15 Can we rely on the market for nding appropriate
solutions for the governance of data in the ecosystem
of connected and automated mobility, or do serious
market failure problems exist that require policy
solutions? This section has the task of identifying
and discussing potential market failure problems
concerning the data governance problem from
an economic perspective.33 Although the policy
discussion about access to data has focused primarily
on the conict between OEMs and independent
service providers, the policy problems regarding the
governance of in-vehicle data and connected cars are
much more complex.
16
For the analysis of this complex data governance
problem, the law and economics of data also have
to be taken into account. Important from a legal and
normative perspective is rst that most of the data in
the connected car are personal data that are subject
to European data protection law, which grants the
data subjects (i.e. the car users) a set of strong rights
in relation to these data in order to protect their
privacy. Therefore OEMs, but also other rms that
would like to use their data, need the consent of
the car users34 for the processing and use of these
data.35 Thus, it is also necessary to discuss whether
the car users are capable of making rational and
well-informed decisions about permitting the OEMs
(or other rms) the use of their personal data, and
whether they are offered sufcient privacy options
33 Kerber/Frank, Data Governance Regimes in the Digital
Economy: The Example of Connected Cars, 2017, available
at: , and Martens/
Mueller-Langer (n 7) seem to be the only papers that have
analyzed this access to in-vehicle data problem from an
economic perspective. The TRL study (n 2)) also takes
economic effects into account, but does not analyze market
failures. For a theoretical framework for the analysis of data
governance problems from an economic perspective see
Kerber/Frank (n 33) 9-17.
34 For some (mostly legal) issues (e.g., data protection), it is
necessary to distinguish between car owners, car drivers,
and car passengers. In this paper we cannot go into the
details of this problem. We therefore will mostly use the
general term ‘car users’ and only in some specic contexts
use explicitly the terms car owners or car drivers, where
this particular role is relevant (e.g. in relation to buying a
car).
35 For the relevance of European data protection law for
data in connected cars, see Hornung, Verfügungsrechte
an fahrzeugbezogenen Daten. Das vernetzte Automobil
zwischen Wertschöpfung und Persönlichkeitsschutz.
Datenschutz und Datensicherheit, 2015, 359-366, Hornung/
Goeble, “Data Ownership” im vernetzten Automobil.
Computer und Recht 2015, 265-273.
for being able to protect their privacy. Secondly,
in the case of non-personal in-vehicle data - which
might be certain kinds of technical data and, in
particular, the huge mass of anonymized data -
no clear legal rights exist, especially no property
rights for data.
36
The discussion about data rights
however, has shown that an exclusive de facto
control of non-personal data by a data holder from
an economic perspective leads to a de facto (but not
legal) “ownership” of these data. But due to the non-
rivalry in the use of data, it is unclear whether such
an exclusive “ownership” of data is an economically
efcient governance solution. Especially in multi-
stakeholder situations, such as the ecosystem of
connected and automated driving, in which the same
in-vehicle data can be used for the value creation of
many service providers, it is very doubtful whether
the exclusive (monopolistic) control of these data
by one stakeholder leads to an efcient way of
using the data.
37
Therefore, the specic economic
characteristics of the data and the data economy are
also an important input for the following analysis
concerning appropriate solutions for the governance
of the in-vehicle data of connected cars.38
17 The analysis in this section is structured as follows.
In section II we will analyze how the technological
decisions of the OEMs - as choosing the extended
vehicle concept or the on-board application platform
- determine who has de facto control of the in-vehicle
data and might therefore be able to appropriate the
benets of these data. Section III offers a critical
analysis of the main argument of the OEMs, that
the extended vehicle concept with its exclusive
control of the access to data and the car is necessary
for ensuring the necessary high level of safety and
security of connected driving. Section IV analyzes
the potential negative effects of the extended vehicle
on competition and innovation on the markets for
aftermarket and complementary services. In section
V it will be shown that competition between OEMs
does not necessarily lead to optimal technological
decisions with regards to interoperability and
standardization leading to a potential market failure
concerning technological choice. This is followed
36 These data might be subject to trade secret protection,
but this does not grant a property-like legal position. See
for this discussion, Zech, A Legal Framework for a Data
Economy in the European Digital Single Market: Rights to
Use Data. Journal of Intellectual Property Law & Practice,
2016, 460-470, and in more detail section D.II.
37 See Kerber (n 4)109-133.
38 See for contributions about the economics of data and the
data economy, OECD, Data-driven innovation: Big data for
growth and well-being, 2015; Kerber, A new (intellectual)
property right for non-personal data? An economic
analysis. GRURInt, 2016, 989-998; Duch-Brown/Martens/
Mueller-Langer, The economics of ownership, access and
trade in digital data. EC JRC Technical Reports Working
Paper 2017-01, 2017; Kerber (n 4) 109-133; Schweitzer/
Peitz, Datenmärkte: Funktionsweise und Regelungsbedarf.
Diskussionspapier 17-043, Mannheim: ZEW, 2017.
Data Governance in Connected Cars
2018
317
3
by an analysis of potential market failures due to
information and behavioral problems of car users
vis-à-vis their consent to the use of their personal
data and the protection of their privacy (section VI).
Section VII offers a brief analysis probing to what
extent these potential market failures might be
mitigated by competition between the manufacturers
of connected cars (section VII). The results of section
C are summarized in the concluding section VIII.
II. Technological decisions
and de facto control of data
and access to the car
18
What are the economic implications of the
technological decision of OEMs for the “extended
vehicle”? Since all in-vehicle data are transmitted
directly to proprietary servers of the OEMs, they are
obtaining de facto exclusive control of these data.
Neither the car users nor other stakeholders can get
access to these data without the consent of the OEMs.
In that respect, the OEMs have gotten the de facto
(but not legal) “ownership” of these data and might
therefore be capable of appropriating the economic
value of these data.39 Additionally, the extended
vehicle concept also implies that the OEMs have
the exclusive control of the access to the connected
car; specically, without the consent of the OEMs,
independent service providers cannot exchange
data with the connected car, nor communicate with
the car drivers via the integrated Human-Machine-
Interface (HMI). Therefore, the connected car is a
closed system (similar to Apple’s iPhone). As far as
the OEMs have exclusive control of in-vehicle data
and the access to the connected car, all independent
service providers who would like to offer services
to the car users need the consent (and therefore
contracts) with the OEMs for being granted access
to: (1) in-vehicle data that they need as indispensable
input for their services; and/or (2) to the connected
car, if they need access either to the IT system or the
HMI of the car for providing these services and/or
communicating with the car users.40 As far as OEMs
39 This de facto exclusive control of these data is only limited
by: (1) the regulated access for repair and maintenance
information (type approval regulation), and (2) by the rights
of the car users regarding their personal data, but these
rights do not extend to non-personal data (and therefore
the anonymized data sets from these personal data).
40 These distinctions are important, because the exclusivity of
the control of the access to the in-vehicle data and the car
by the OEMs is limited by alternative channels for getting
data and/or for communication with the car users (as, e.g.
through smartphones). Therefore, e.g., location data and
data about the trafc situation might not be exclusive,
because this information might also be obtained from the
smartphones of the car users or from connected cars from
other brands. The importance of the number of data access
channels is emphasized by Kerber/Frank (n 33), 41 and
have exclusive control, the consumers can also only
choose between those service providers who have
contracts with the OEMs. Since the connected car
is an expensive durable good, the car owners are
“locked in” the closed system of the OEMs. Therefore,
the OEMs are in a “monopolistic” gatekeeper
position with regards to the in-vehicle data41 and
the connected car and can increase their prots by
“selling” access to the users of the connected car to
the independent service providers.
19
Technological alternatives would lead to different
data governance solutions. The “on-board application
platform” – the technological architecture favored
both by the TRL study and independent service
providers42 – would offer the possibility that car
users can decide where the data are stored and
whom they grant access to the in-vehicle data and/
or the connected car. Therefore, it would be the car
users who have the exclusive control. In this case
they can choose freely between all service providers
without the need to have contracts with the OEMs.
As a consequence, the car users would be the de
facto “owners” of these data and can use them for
their own benet, either through choosing the
most attractive offer from service providers and/
or by “selling” these data to the highest bidder.
With this technological solution, the OEMs would
have lost their “monopolistic” gatekeeper position
regarding in-vehicle data. Hence, from an economic
perspective the technological solution determines
the initial allocation of the de facto exclusive control
of data and thus the initial allocation of the de facto
“ownership” of data.43 It also decides to what extent
the connected car is a “closed” or an “open” system,
i.e. whether the manufacturer of a primary product
(here the connected car) does also exclusively control
the access to the connected car for aftermarket and
complementary service providers, and whether and
to what extent the consumers are “locked in” (see
below section IV).44
Martens/Mueller-Langer (n 7) 24.
41 See also Martens/Mueller-Langer (n 7) 14.
42 See section B.
43 See Kerber/Frank (n 33) 28. It is important however to
take into account that through contractual arrangements
between the car owners and the OEMs this position
of exclusive control of data and therefore the de facto
“ownership” can be traded between the contracting parties;
see also section C.VI.
44 See Determann/Perens (n 13).
2018
Wolfgang Kerber
318
3
III. Justification of the “extended
vehicle” through safety
and security concerns?
20
The OEMs defend the extended vehicle concept with
the argument that only through this technological
solution (with an external server) a maximum
standard of safety and security can be ensured.45
There can be no doubt that safety and security issues
are very important when it comes to connected and
(automated) driving, especially for the car users. In
the current policy discussion, the problem of access
to in-vehicle data and resources has primarily been
seen as a trade-off problem between safety and
security on the one hand, and fair and undistorted
competition on the other hand. However, it can be
asked whether and to what extent such a trade-off
exists. We will analyze this problem in two steps.
21 In a rst step, we ask whether the external server
solution (as part of the extended vehicle concept)
and the on-board application platform can solve the
safety and security problem. This is a technological
question that has to be answered by technical and
IT experts. The TRL report came to the conclusion
that both the on-board application platform and the
external server solution can solve the safety and
security problems, although there might be cost
advantages for the external server solution.46 Among
IT experts there is a wide-spread opinion that closed
proprietary systems need not be more secure than
well-designed open systems; on the contrary, the
often multi-layered architecture of interoperable
open systems might even offer better protection
against cybersecurity attacks.47 Since OEMs also
offer direct access to their connected cars to some
service providers with whom they have contractual
arrangements, ensuring a sufciently high level
of safety and security seems to also be possible for
direct access to the connected car. However, it is clear
that an open interoperable telematics (on-board
application) platform need the implementation of a
sophisticated safety and cybersecurity system. One
part of the solution might be the separation of safety-
and security-sensitive functions and data from the
vast amount of other data, which are not related to
safety and security.48 It is particularly important,
however, to strictly control whether independent
service providers who want to offer their services
to the car users fulll certain standards for safety
45 Safety and security refer to the safety of the car but also
to cybersecurity of the connected car, which also can
encompass the security of the personal data, see ACEA
2016b (n 15) 5.
46 See TRL (n 2) 77.
47 See in more detail TRL (n 2) 75-79; Determann/Perens (n 13)
939-942, and Martens/Mueller-Langer (n 7) 12.
48 This could be achieved by using so-called hypervisor
technologies (TRL [n 2] 8).
and security of their services; for example, when
it comes to apps and software that are uploaded
to the connected car. This can be implemented by
requiring a certication of these service providers.49
In addition to that, the medium- and long-term
development of integrated mobility systems with
connected, automated, and later autonomous cars
would require in any case the development of a
comprehensive safety and security architecture
with interoperable brand-independent industry-
wide interfaces between connected cars and other
entities. Therefore, solving the safety and security
problems of interoperable telematics platforms, by
for example, establishing a comprehensive system
of certications for safety and security, is in any case
one of the important tasks for achieving the policy
objective of a future integrated mobility system of
connected and automated driving.50 51
22
Most important for the governance of the in-vehicle
data is, however, that safety and security concerns
do not lead to a justication for the exclusive
economic control of the in-vehicle data through the
OEMs.
52
Even if we assume that it is necessary that all
data have to be transmitted to an “external server”
and the OEMs must have exclusive control of the
access to the IT system of the car due to safety and
security reasons, this does not lead to a justication
that they also need to be the de facto exclusive
“owners” of these data with the right to exploit these
data commercially. With regard to connected cars,
the OEMs can also be seen as service providers of IT
security who have the task of keeping the car and its
data safe and secure, whereas the car users still retain
the right to decide who should get access to the in-
vehicle data of the car or to “sell” these data to other
rms. Therefore even if safety/security problems
make it necessary that the OEMs exclusively control
the access to the car and the data, it is not clear at all,
why the OEMs should also have the right to decide
freely and according to their own interests who
can get access to the car and/or the data, and who
can exploit these decisions about access to increase
their prots. The extended vehicle concept thus
entails a bundling of the task of providing safety
and security services with the transfer of de facto
ownership rights of the data to the OEMs, which is
not necessary and lacks economic justication. These
49 Certication was also the regulatory solution concerning
solving quality concerns with regards to the products
of independent spare part producers in the automotive
industry.
50 See EU Commission 2018 (n 1) 9.
51 Particularly important is the solving of liability problems.
See also, Determann/Perens (n 13) 984-986, concerning the
general problems with liability in the case of open systems.
This is however no serious argument against interoperable
telematics systems. If the safety and security problems can
be solved, then also suitable solutions for the assignment of
risks in tort law can be found.
52 See also Kerber/Frank (n 33) 54.
Data Governance in Connected Cars
2018
319
3
are two different roles that can easily be separated
and unbundled.53 One simple “unbundling” solution
in the case of an “external server” solution is the
already much discussed “shared server” solution, in
which the external server is not under the exclusive
control of the OEM but under the control of an entity
that is independent from the OEMs. This entity then
can give access to these data on a non-discriminatory
basis according to certain general principles (e.g.
FRAND conditions) and would therefore eliminate
the privileged position of the OEMs vis-à-vis the data.
In the case of the “on-board application platform”,
it is clear that it is the car user who has de facto
“ownership” of the data and the right to decide on
the access to the car, and the OEMs are “only” service
providers for the safety and security of the car.
23
Therefore, this section leads to the following results:
(1)
The exclusive (“monopolistic”) control of the
in-vehicle data in the extended vehicle concept
that allows the OEMs to appropriate the
economic value of the data cannot be defended
through safety and security concerns. Even
if exclusive control of the access for solving
safety and security problems is necessary, this
does not imply that the provider of safety and
security also needs to have the right to exploit
the commercial value of the data. Both roles can
be easily unbundled.
(2)
However, it is also very doubtful whether an
external server solution and the car as a closed
system with the exclusive control of the OEMs
concerning the access to the car is necessary
at all for safety and security. There seem to be
good reasons to believe that the same (or even
a higher) level of safety and security can also
be achieved by using an “on-board application
platform” with a sophisticated safety and
security system.
(3)
As a consequence, the basic assumption of
the current policy discussion that there is a
fundamental trade-off between the objectives
of safety/security and fair and undistorted
competition is deeply awed. There is denitely
no such trade-off with regards to the access to
the in-vehicle data, and it is also very doubtful
whether there is such a trade-off related to
access to the connected car.
53 A rm who hires a security service rm for the task
to control the access to this rm (either physically or
concerning its IT system) does not simultaneously give
the security rm the right to decide freely whom to give
access to the rm and whom not, and therefore allowing the
security rm to “sell“ access to this rm. The right to decide
who gets access will remain with the rm. The security rm
has only the right (and duty) to deny access in the case of
clearly dened safety and security risks.
(4)
Another conclusion is that it is necessary to
analyze the safety and security problems as part
of the medium- and long-term technological
architecture of an integrated ecosystem of
connected and automated mobility (see below
section C.V.).
IV. Competition problems
on aftermarket and
complementary markets
24
In section B we have seen that both the independent
service providers and the consumers are very
concerned that the exclusive control of the OEMs
regarding the data and access to the connected
car can impede competition and innovation on the
markets for aftermarket services and complementary
services in the ecosystem of connected driving.
The problem of ensuring fair and undistorted
competition for independent service providers has
been raised in the Working Group 6 of the C-ITS
platform, conrmed by the TRL study, and has
been acknowledged by the EU Commission as an
unsolved problem.54 From a competition economics
perspective, the competition concerns have to be
taken very seriously. As far as independent service
providers need access to in-vehicle data and/or the
access to the connected car, the OEMs can control
a necessary (“essential”) resource for providing
these services. This position allows them to foreclose
independent service providers. This is an old well-
known competition problem in the automotive
industry,55 and the long-existing regulatory efforts
of European competition policy for protecting
competition on markets for automotive repair and
maintenance services and spare parts, which led to
the solution of a “regulated access” to necessary
technical information (see section B), have always
focused on exactly this problem. Since many more
new and innovative services are expected to be
offered in the context of connected driving, the
problem of foreclosing competition and leveraging
market power has gotten much more important than
in the traditional case of repair and maintenance
services. It is also important that the problem is
not limited to automotive aftermarket services, but
also encompasses the wide range of many other
innovative services for the users of connected cars,
54 See EU Commission 2018 (n 1) 13.
55 For the economics of aftermarkets and its discussion in
competition law see, Shapiro, Aftermarkets and Consumer
Welfare: Making Sense of Kodak, Antitrust Law Journal,
1995, 483-511; Borenstein/MayKie-Mason/Netz, Exercising
Market Power in Proprietary Aftermarkets, Journal of
Economics & Management Strategy 9, 2000, 157; Bauer,
Antitrust Implications of Aftermarkets, Antitrust Bulletin
52, 2007, 31, and Bishop/Walker, The Economics of EC
Competition Law, 2010, 150-152, 245-249.
2018
Wolfgang Kerber
320
3
such as complementary services,
56
which are also
often the result of new data-driven innovation.
25 The exclusive control of the data and the car allows
the OEMs several options for increasing their prots
through this gatekeeper position. One option is to
deny access in order to block the entry of service
providers for specic kinds of services, which then
could be offered exclusively by the OEMs themselves.
If these markets promise particularly high prots,
then monopolizing these markets can be one
strategy for making prots through foreclosure
strategies. Another option is to “sell” access to these
data and the car to independent service providers
who would like to enter these markets. This can be
done by concluding B2B-agreements with service
providers who, for a certain price, can get access to
data and / or the IT system of the car, which can be
interpreted as an entry fee into the relevant markets.
This can also lead to exclusivity agreements; namely,
that such a “license” to sell services in the connected
car is granted only to one service provider for a
high “fee” that allows the OEMs to reap the prots
from such an exclusive position of providing a
specic service for the cars of a particular brand.
But even if the OEMs grant access to a number of
service providers, the OEMs remain in control of
the aftermarkets and complementary services via
their contractual relationships with these rms.
Irrespective of the option the OEMs choose for
maximizing their prots,57 there are no independent
markets for aftermarket and complementary
services any more, and the OEMs can reap all (or
most) of the prots. Moreover, the concern that
such market control can lead to less innovation of
new services has to be taken very seriously from
an innovation economics perspective, because it
enables the OEMs to lter which innovative services
are offered to the car users. An additional way of
monetizing the data is the selling of (anonymized)
data sets for all kinds of other uses outside of the
automotive industry and the ecosystem of connected
and automated driving. Since many of these data sets
are unique and not replicable, there is a danger that
the ensuing monopolistic prices will lead to welfare
losses through an under-utilization of these data in
the data economy.58
56 The term “complementary services” encompasses all
services that are useful for the car users only in connection
with the connected car, especially during driving. Therefore,
the car and these services are economically complements.
In that respect, there is no difference between aftermarket
services and other complementary services from an
economic perspective.
57 Please note that the OEMs with their extended vehicle
concept insist on freely negotiated B2B agreements (ACEA
2016a [n 15]), i.e. that it is in their discretion what kind of
prot-maximizing strategy they use.
58 See Martens/Mueller-Langer (n 7) 14-17, who also analyze
pricing strategies of OEMs for selling access to data
(monopoly pricing, price discrimination).
26 Furthermore, other variants of the external server
solution have been discussed. One variant is that
“neutral servers” - operated by independent entities
- might be established, which provide in-vehicle data
to other stakeholders under non-discriminatory
terms. This neutral server solution however, suffers
from the problem that the in-vehicle data are still
rst transmitted exclusively to a proprietary server
of the OEMs, who are free to decide what data
they make available under what conditions in free
B2B-agreements to the operators of these neutral
servers. Therefore, the OEMs can still apply the same
strategies as described in the last paragraph. The
only difference is that the OEMs cannot make direct
contracts with the users of those data that are made
available to the neutral servers, which limits their
options for controlling the use of these data to some
extent.59 Whereas such a neutral server solution is
not a solution for the competition problems, this is
different for the already mentioned “shared server”
solution. Since in this case the in-vehicle data are
transmitted directly to an external server operated
by a neutral entity, the OEMs lose their monopolistic
gatekeeper position regarding in-vehicle data.
This leads to a level playing eld with regards to
the access to the data, and therefore removes one
important hurdle for ensuring fair and undistorted
competition on the markets for aftermarket and
complementary services. However, a shared server
would not necessarily solve all competition problems
on these markets, because the OEMs might still block
independent service providers via their exclusive
control of the access to the car. A transition to an
open on-board application platform might also solve
this problem.
27
Therefore, from a competition economics
perspective, there can be no doubt that the OEMs can
eliminate competition on markets for aftermarkets
and complementary services due to their exclusive
control of the in-vehicle data and the access to the
car. In that respect, the concerns of the independent
service providers about the implications of the
extended vehicle concept are justied. However,
59 One benet of this neutral server solution can be that it
might help to mitigate the concern of the independent
service providers, that by monitoring their proprietary
server the OEMs can observe the transactions between
car users and independent service providers, which might
give them an advantage regarding the offering of their
own services. This is a wide-spread concern of independent
service providers. See C-ITS Platform (n 2) 79. Please note
that the same competition problem is discussed currently
in the context of transaction and user data on platforms
such as Amazon. Here the concern is that those platforms
can use these data for favoring their own services (see
Schweitzer/Haucap/Kerber/Welker, Modernisierung der
Missbrauchsaufsicht für marktmächtige Unternehmen,
2018, 142), as well as the current Amazon investigation of
the EU Commission (see
de/amazon-investigated-by-eu-commissioner-margrethe-
vestager-2018-9?r=US&IR=T>).
Data Governance in Connected Cars
2018
321
3
from an economic perspective an important
counterargument has to be considered. It also should
be asked whether competition between the OEMs is
capable of solving the problem of ensuring an efcient
provision of aftermarket and complementary
services and with prices on a competitive level,
even if the OEMs have exclusive control of these
markets. Competition between OEMs can also be
seen as competition between connected cars as
bundles of the car itself and a set of aftermarket and
complementary services (“system competition”). It
can be argued that if competition between OEMs
works very well, then they might be under enough
competitive pressure for offering attractive bundles
of cars and services at competitive prices. Otherwise
car buyers would switch to the connected cars of
other brands. This is a standard argument in the
economic theory of aftermarkets. This question
has also emerged in competition law with respect
to dening the relevant markets in the automotive
industry. Is the relevant market an aftermarket for
a specic brand because consumers are “locked in”
after they bought a particular car? This would lead
to the conclusion that an OEM is a dominat rm in
regard to aftermarkets and complementary services,
which depend on the access to the data or the car. Or
do the car buyers decide between different bundles
of cars and services of OEMs leading to the denition
of “system markets”?
28
Can competition between bundles of OEMs and
aftermarket and complementary services work
well enough for solving the problems of exclusive
control of OEMs? This problem has been discussed
in competition economics extensively,60 for example
in the context of the well-known printer/toner
problem. If we assume that the consumers are
rational and well-informed about the future costs of
the specic toner they need before buying a printer,
then the ensuing result that the buyers are getting
locked-in regarding the toner is no problem, because
they would already have taken this into account in
their decision to buy the printer. However, even in
this relatively simple lock-in problem, consumers
seem to have considerable problems in dealing with
it. These problems are much larger for the car buyers
in the case of connected and automated mobility. It
is very hard for car buyers to make reliable estimates
about the future costs of being locked into such a
bundle. The car buyers cannot know what kinds of
services with what prices the OEMs will offer during
the lifetime of a connected car. In the same way, they
will not know what kind of choice between different
service providers the OEMs will offer them in two,
ve, or eight years.
61
Therefore it is very doubtful
60 See Shapiro/Teece, Systems Competition and Aftermarkets:
An Economic Analysis of Kodak, Antitrust Bulletin, 1994,
Shapiro (n 55), Borenstein et al (n 55), Bauer (n 55), Bishop/
Walker (n 55) 150 et seq., 249 et seq.
61 Selling the connected car in the case that OEMs later
whether the car buyers can appropriately calculate
the long-term costs and benets of the aftermarket
and complementary services that are part of
this bundle. As a consequence, it is very unclear
whether system competition between OEMs can
work sufciently for solving the competition and
innovation problems on the markets for aftermarket
and complementary services. It should be noted that
if system competition between OEMs would have
worked effectively in the past, the decades-long
efforts in competition law for protecting competition
in the markets for repair and maintenance services
(as well as spare parts) would not have been
necessary. Since connected and automated cars
are much more complex in regard to services than
traditional cars, we should be very cautious in
relying on the effectiveness of systems competition
between OEMs in regard to these services.
V. Market failures in regard
to technological choice:
interoperability and
standardization problems
29
In economics we usually assume that the rms
should be free to decide on the technological design
of their innovations and that the market is capable
of selecting the superior technologies. If the OEMs
choose the extended vehicle concept and this
solution also prevails in the markets (as it is widely
expected without regulatory intervention),62 the
question arises whether it is also the most efcient
technological solution or whether there might be a
market failure problem about technological choice.
The TRL study came to the conclusion that in the
long-term the on-board application platform would
be superior to the extended vehicle concept (with its
external server)63 and also our analysis will suggest
a similar result. Economic research has identied a
number of cases, in which prot-maximizing rms
can choose inefcient technologies and/or markets
are not capable of selecting the best technologies.64
Since in the future ecosystem of connected and
automated mobility, interconnectivity and real-
diminish the choice or increase prices for these services is
not a solution, because this will lead to lower prices of the
used cars.
62 See TRL (n 2) 13.
63 See TRL (n 2) 170.
64 In one group of cases dynamic economies of scale (learning
effects) or network effects can lead to path dependencies
which might result in the lock-in of old technologies which
are hard to be replaced with newer more efcient ones.
The famous QWERTY-problem is another example. See,
e.g., Katz/Shapiro, Network Externalitites, Competition
and Compatibility, American Economic Review 75, 1985,
424; David, Clio and the Economics of QWERTY, American
Economic Review 78, 1988, 332.
2018
Wolfgang Kerber
322
3
time exchange of data between cars, infrastructure,
private rms, and public institutions will be necessary
for a well-functioning integrated mobility system,
interoperability and standardization are important
issues in this mobility ecosystem. Therefore, it can
be asked whether there could be potential market
failure problems with regards to interoperability and
standardization.
30
One interoperability issue refers to the question
whether OEMs choose a proprietary and closed
technological system for the connected car or an
open interoperable system, in which the car users
can decide about the access to the connected car.
The economics of interoperability shows that both
open and closed systems can have benets and costs,
and that a deeper economic analysis is necessary for
answering the question of which one is superior in a
specic case.
65
Our discussion about the effects of the
extended vehicle concept vs. the on-board application
platform can be seen as part of such an assessment of
the advantages and costs of interoperability in the
case of connected driving. Since one of the benets
of interoperability can be more innovation, the
question arises whether a closed system would lead
to more innovative solutions (e.g., due to synergies
between the connected car and other services within
the system) or whether it can be expected that,
due to open interfaces, an open system that allows
for independent innovation activities of service
providers would lead to more innovative services
within the ecosystem of connected and automated
mobility. So far, the OEMs have not claimed that
their closed systems will lead to more innovation
in aftermarket and complementary services,
whereas the independent service providers are
emphasizing the huge potential of new innovative
services. Therefore, it is necessary to carry out
a much deeper analysis of the advantages and
costs of interoperability to decide which degree of
closeness or openness of the connected car would
be optimal. If competition between entire systems
(bundles of cars and services) does not work well
with connected cars, as we suggested in the previous
section, then it is doubtful whether individual prot-
maximizing decisions of the manufacturers of the
primary products (here: the connected cars) lead to
65 On the one hand, more open systems with more
interoperability can offer the consumers more choice,
innovation and competition between complementary
products and services that they can use in combination
with this system. On the other hand, closed systems might
have advantages in terms of more differentiation and
a higher quality of services due to a better integration
between the system and these complementary services.
See for the economics of interoperability Choi/Whinston,
Benets and requirements for interoperability in the
electronic marketplace, Technology in Society 22, 2000, 33;
Gasser, Interoperability in the Digital Ecosystem, 2015, 9-17;
available at: , and as
overview Kerber/Schweitzer, Interoperability in the Digital
Economy, JIPITEC, 2017, 39, 41 et seq.
optimal decisions about interoperability regarding
complementary services. Rather the rms tend to
choose a proprietary closed system too often.66
31
However, interoperability is also very relevant at the
level of the entire integrated ecosystem of connected
and automated mobility. Due to the long-term
need for direct communication and data exchange
between vehicles, infrastructure, private rms and
public institutions, far-reaching standardization
processes concerning communication, data formats
and categorization, safety and cybersecurity issues
and other technological features are necessary,
which require industry-wide standardized interfaces
between the vehicles and the overall technical
architecture of the mobility system. The connected,
automated and later autonomous car must be an
integral part of this system, specically, the cars
have to t into the overall architecture and therefore
have to comply with standardized technical
interfaces in order to be capable to interoperate with
many other parts of this ecosystem. Therefore an
(to some extent) open and interoperable on-board
application platform has to be developed in any case
in the next steps of the automation of the connected
cars.
67
The economics of standard-setting has shown
that these kinds of uniform standards at the level of
the entire mobility system cannot emerge in market
competition.68 Although the decisions of the OEMs
for the extended vehicle concept might seem to be
prot-maximizing - at least in the short- or medium-
term - in such situations their individual incentives
might lead them to technological decisions that are
not optimal for the entire ecosystem. Therefore, it
is necessary to nd a solution for this market failure
problem. This can be done by a collaboration of all
relevant stakeholders in this ecosystem in order to
develop the most suitable technological standards
and interfaces.69
66 For the general complaint in the digital economy about too
many proprietary solutions and not enough interoperability
see, e.g., PwC, Cross-cutting Business Models für IoT. Final
report (SMART number 2017/0027), Brussels, 2017, 132.
67 See Martens/Mueller-Langer (n 7) 13 regarding the
necessity of on-board application platforms for automated
and autonomous driving.
68 Due to the advantages of compatibility, often only one
uniform (and monopolistic) standard can exist. In the
economics of standard-setting it has been shown that
markets encounter large problems when it comes to nding
and establishing efcient standards in an uncoordinated
way. The main problem is that prot-maximizing individual
rms often have incentives to choose technological
standards that are not aligned with the overall welfare
effects of these standards. Due to these market failure
problems, many standards are developed through standard-
setting organizations (SSO), in which rms collaborate to
create new standards. For an overview about the economics
of (the market failure problems of) standard-setting see,
Farrell/Simcoe, Four Paths to Compatibility, in: Peitz/
Waldfogel, The Oxford Handbook of the Digital Economy,
2012, 34-58.
69 Efforts for standardization for improving interoperability
Data Governance in Connected Cars
2018
323
3
VI. Information and privacy
problems of consumers
32
The discussion on the governance of in-vehicle
data has been dominated by the conict between
the OEMs and independent service providers about
access to in-vehicle data. Much less attention has
been paid to potential market failures concerning
the interests of the consumers, i.e. the car users.
First it is important to understand that buying a
connected and automated car requires not only a
traditional sales contract but also contracts about
services (and software updates etc.), as well as
contractual provisions about the consent of the car
users for the processing and the use of personal data
in the connected car. Therefore, both parties are de
facto in a long-term relationship, which implies a
much larger “lock-in” problem for the car owners
than for traditional cars. This “lock-in” problem
does also exist in the solution of the “on-board
application platform” but is much more serious in
the “extended vehicle” concept, where the OEMs
also can control additionally many aftermarket and
complementary services and the consumers are
“locked-in” in the entire bundle of car and services
(see section C.IV.). However, in the following, we
want to focus on the problem whether there might
be a market failure problem when giving consent for
using personal data and the protection of privacy.
The following reasonings refer again mainly to the
extended vehicle concept.
33
In the discussion about privacy problems in the
digital economy and the issue of “data as counter-
performance” for “free services” as in the case of
the Google search engine or social media (Facebook),
serious concerns have been raised, whether the
“notice and consent” solutions in standard form
contracts for giving digital companies permission
to use their personal data work in a satisfactory
way.
70
This refers to the problem of transparency
regarding the extent of data collection and the use
of the data, whether users are aware of the value of
their data, as well as the problem of whether there is
a real choice if without giving consent these services
cannot be used. Related to that, also whether enough
privacy options are offered, i.e. that users can make
granular decisions about providing personal data
according to their specic privacy preferences. All
of these problems are also relevant in relation to the
are already taking place both at the EU and the international
level. See TRL (n 2) 58-67, and EU Commission 2018 (n 1) 4-8.
70 See, e.g., European Data Protection Supervisor, Privacy
and competitiveness in the age of big data: The interplay
between data protection, competition law and consumer
protection in the Digital Economy, Preliminary Opinion,
2014; Borgesius, Behavioural Sciences and the Regulation
of Privacy on the Internet, in: Alemanno/Sibony, Nudging
and the Law – What can EU Law learn from Behavioural
Sciences?, 2015, 179-207.
personal data of the connected cars. In the context of
the “privacy paradox” discussion it has been argued
that due to information and behavioral problems,
users might often not be capable of making rational,
well-informed decisions about providing personal
data and protecting their privacy.71 Therefore, it
is necessary to conduct further research into the
contractual arrangements between car owners and
OEMs vis-à-vis the provision of personal data and the
possibilities for protecting their privacy, and also ask
whether also in this context market failure problems
and unsolved privacy problems exist.72
34
In the current policy discussion surrounding data
governance in connected cars, there is a consensus
that the privacy of the car users has to be protected.
However, so far not much specic discussion can
be found regarding how this should be achieved.
In a recent survey, car owners in the EU were very
concerned about disclosure and commercial use of
personal data in connected cars, and emphasized
their wishes for the ability to make more granular
decisions about the provision of personal data.
73
This
can imply that car users do not have to generally
give consent to the processing and use of personal
data, but that, for example, car users can decide for
each ride whether location data are transmitted
or not. The experiences with the privacy policies
in other digital contexts do not support the belief
that competition between OEMs might be enough
in order to lead to privacy-friendly solutions for
car users. Therefore, a discussion about additional
regulatory solutions (perhaps also in the form of
self-regulation) might be necessary for supporting
privacy-by-default solutions and offering sufcient
choice between different privacy options.74
71 For the privacy paradox, and the (behavioral) economics
perspective see, Norberg/Horne/Horne, The privacy
paradox: Personal information disclosure intentions versus
behaviors. Journal of Consumer Affairs, 2007, 100-126,
Kokolakis, Privacy attitudes and privacy behavior: A review
of current research on the privacy paradox phenomenon.
Computers & Security, 2015, 122-134, Hermstrüwer,
Contracting around privacy: The (Behavioral) Law and
Economics of Consent and Big Data. JIPITEC, 2017, 9-26,
and Acquisti/Wagman/Taylor, The Economics of Privacy,
Journal of Economic Literature, 2016, 479.
72 See also Metzger, Digitale Mobilität - Verträge über
Nutzerdaten, forthcoming in: GRUR 2019 (2). For a skeptical
view about individual consent with regards to protecting
privacy in connected cars, see from a U.S. perspective,
Akalu, Privacy, consent and vehicular ad hoc networks
(VANETs). Computer Law & Security Review, 2018, 37.
73 See FIA 2016b (n 22) 15.
74 In that respect also, the discussion about Personal
Information Management Systems (PIMS) may be relevant.
See European Data Protection Supervisor, EDPS Opinion
on Personal Information Man-agement Systems. Opinion
9/2016.
2018
Wolfgang Kerber
324
3
35
Another very interesting question is whether the
car users should also have rights concerning the
non-personal data of their cars, especially, also the
anonymized sets of (their) data, and to what extent
they get a (fair) share of the value of the data of
their connected car.
75
This is a very difcult problem
that cannot be analyzed here in detail. Therefore
only a few remarks can be made. There is a wide-
spread opinion that the owner of a car should also
“own” the data which are produced in the car
(“MyCarMyData”). 76 However, from an economic
perspective, it should be taken into account that
providing data to the OEMs can be seen as an
example of “data as counter-performance” as part
of the contractual arrangements between OEM and
the car owners, which from an economic perspective
might lead under competitive conditions on the car
market to lower prices for the car and its services. In
this case the car owners might indirectly participate
in the value of the data. However, it also has to be
taken into account whether this mechanism really
works sufciently.
77
All of these questions require
much more research. They also arise to some extent
with the technological solution of the “on-board
application platform” solution; however, in this case
the car users could also “sell” their data directly to
other rms than the OEMs.
VII. Can competition between
car manufacturers solve the
market failure problems?
36
Competition between OEMs can only have a very
limited effect on market failures through information
and behavioral problems of consumers when giving
their consent to the provision of data and protecting
privacy. Competition between OEMs can also not
solve the potential market failure problems in the
case of choosing the optimal technologies concerning
technical standards and interoperability with
regards to an optimal technological architecture for
an integrated ecosystem of connected and automated
mobility. Since there are good reasons to be skeptical
about the effectiveness of systems competition
between OEMs, it also cannot be expected that this
competition would solve the competition problems
on the market for aftermarket and complementary
services that are caused by the exclusive control of
the access to the in-vehicle data and the car in the
extended vehicle concept. But from a competition
economics perspective, the question of the impact
75 See e.g., BEUC (n 23) 8; Specht/Kerber (n 2) 190.
76 See FIA 2016b (n 22) 1.
77 For the problem of whether the provision of data to the
OEMs would lead to lower prices for connected cars, see
Kerber/Frank (n 33) 28.
of competition between OEMs is very important and
requires much more research. In that respect it is
also important that competitive pressure on the
OEMs can also come from outside the automotive
industry. Large digital companies such as Google,
Apple, and others, also want to enter this ecosystem
of connected and automated mobility, either with
their own connected and automated cars, or with
their huge competence concerning data analytics
and articial intelligence and the provision of
many digital services.
78
Especially strategic alliances
between traditional OEMs and large digital companies
have the potential to intensify competition between
OEMs and might break up the old business model
of the OEMs. Therefore, a careful monitoring of the
business strategies of the OEMs is important.
37
However, there might also be competition
problems between OEMs through collusive, cartel-
like behavior of the OEMs. It can even be asked
whether the extended vehicle concept itself - as it
has been developed by OEMs and defended by their
associations in Europe - can be seen as an anti-
competitive horizontal agreement about decisions
on technology and governance of in-vehicle data
in connected and automated cars. All OEMs that
apply the extended vehicle concept (1) use the
same technological solution of a proprietary server
to which all in-vehicle data are transmitted (leading
to their exclusive control of the in-vehicle data), and
(2) design the connected car as a closed system (with
exclusive control of the access to the car). Therefore,
the monopolistic gatekeeper position of the OEMs
is an integral part of the extended vehicle concept.
It would also be interesting to investigate to what
extent the OEMs with the extended vehicle concept
have also agreed upon (3) the categories of data that
they are making accessible under certain conditions
to other stakeholders, and (4) on contractual
provisions concerning (personal) data and privacy
options in their contracts with car owners. As far as
the OEMs have de facto agreed on these and perhaps
also other aspects of their technological or data
governance solutions, competition in regard to these
solutions would have been eliminated.
79
There would
78 Martens/Mueller-Langer (n 7) 20-23 make the important
argument that if platforms such as media and entertainment
platforms with large network effects offer car versions
of their (for the car users very attractive) services (Apple
iOSCarPlay or Android Auto), then OEMs might be under
competitive pressure to install those media systems in their
cars as part of the entire bundle they are offering to their
customers. This would allow the large digital companies
to enter the markets of aftermarket and complementary
services and use their huge competitive advantages with
regards to data and data analytics on these markets.
79 In that respect also, a closer analysis of the effects of the
standard-setting process in regard to the “Extended
Vehicle Standard” (ISO 20078) might be relevant (for more
information see,
html>). In regard to technological collusion between
OEMs in the automotive industry see also the current
Data Governance in Connected Cars
2018
325
3
be, in particular, no competition regarding other
technological solutions such as the interoperable
on-board application platform. Since, however,
the business strategies of OEMs also differ to some
degree,
80
it would be necessary to investigate the
extent to which the extended vehicle concept leads to
collusion between the OEMs in regarding the design
of technological and data governance solutions in
the ecosystem of connected and automated driving.81
VIII. Conclusions
38
In this section we have analyzed what kind of
market failure problems might emerge concerning
the data governance in the ecosystem of connected
and automated mobility and offered a preliminary
assessment of these market failures, which however
requires much more (and primarily empirical)
research:
(1) Competition problems: By using the extended
vehicle concept with its exclusive control of
the access to the data and the car, the OEMs
can foreclose independent service providers
and control and monopolize aftermarket and
complementary services. This can lead to too
high prices, not enough consumer choice, and
less innovation. These competition problems
cannot be sufciently mitigated through
systems competition between OEMs.
(2)
Interoperability and standardization
problems: Within this complex integrated
ecosystem of connected and automated mobility,
it cannot be expected that the individual prot-
maximizing decisions of OEMs on technology
lead to optimal solutions when it comes to
interoperability and standardization for the
entire system.
(3)
Information and privacy problems of car
users: Especially important is research into
whether and to what extent there might also
be a market failure problem regarding the
decisions of the car users to give their consent
to the processing and use of their personal
data. This would also require an analysis of the
provisions on data in (standard form) contracts
and the options the OEMs offer the car users
for granular decisions about protecting their
privacy.
investigation of the EU Commission into possible collusion
on clean emission technology (see press release IP/18/5822,
18 September 2018).
80 See TRL (n 2) 67-72.
81 Then the question of a cartel exemption can be discussed
(see below section D.III.).
(4) Safety and cybersecurity: These concerns are
very important but do not lead to a justication
of the extended vehicle concept, because they
can also be solved with the on-board application
platform. In any case, safety and cybersecurity
concerns cannot justify the exclusive control
and therefore de facto ownership of the in-
vehicle data by the OEMs.
39 What are the conclusions for the current discussion
between OEMs and the independent service
providers about access to in-vehicle data? Although
there is still considerable need for further research,
the preliminary results of our analyses of potential
market failure problems suggest that the concerns
of the independent service providers regarding
the impact of the extended vehicle concept on
competition and innovation on the markets for
services in the ecosystem of connected driving are
justied. Since the extended vehicle concept with
its exclusive control of the in-vehicle data cannot be
defended by safety and cybersecurity concerns, the
trade-off between competition and cybersecurity
does not exist in relation to in-vehicle data. Safety
and security concerns also seem to be solvable with
the on-board application platform, which would
allow the provision of control of the access to the
connected car and the in-vehicle data to the car
users. Both the “shared server” in the case of the
current technological solution of the “external
server” and the on-board application platform would
allow for a “level playing eld” in terms of the access
to in-vehicle data, and can therefore contribute to
the protection of competition on the markets for
services within the ecosystem of connected and
automated mobility.
D. Governance of in-vehicle data:
Discussion of policy approaches
I. Complexity of the data
governance problem
40
Although the conclusions in the last section seem
to support the position of the independent service
providers, the data governance problem in this
ecosystem of connected driving is much more
complex. Whereas both the “shared server” and the
“on-board application platform” offer the chance
to eliminate the exclusive control of the OEMs
regarding the in-vehicle data, they are themselves
neither a clear nor a comprehensive solution for
the governance of the in-vehicle data. There are
many open questions; namely, who should operate
a shared server and how should it grant access to
what kinds of data, and under what conditions?
2018
Wolfgang Kerber
326
3
Should all data that are produced in the car be
transmitted to this server, or do OEMs and, e.g.,
component suppliers have direct access to certain
kinds of technical data (safety and cybersecurity
reasons, business secrets)? How should one deal with
data that are costly to produce compared to those
with negligible costs? Should there be one shared
server for each OEM or might it be better to pool
the in-vehicle in one industry-wide shared server
for a better exploitation of the advantages of data
aggregation? Also, the proposal to transition to an
interoperable on-board application platform does
not clarify how the governance of the in-vehicle
data will look like under this technological solution.
These policy proposals also do not take into account
the potential market failure problems when it comes
to information and privacy problems of car users
concerning the provision of personal data and the
protection of their privacy. They also do not consider
the question of whether and how car users should
participate in the value of the data. In addition to
that, there may be many more proposals for solving
the problems, such as voluntary measures like
principles for the access to data.
41
These questions should only emphasize that the
data governance problem in the ecosystem of
connected and automated mobility is a very complex
problem that requires much more research from a
technological, economic and legal perspective.
82
This
paper does not claim to have a clear policy proposal
about the governance of these data, although it
clearly suggests that the currently existing extended
vehicle concept is not a suitable concept and that
it is therefore necessary to think about (perhaps
far-reaching) policy solutions. In the following,
we will present an overview about some current
policy discussions regarding the governance of data
and ask to what extent they might be helpful for
solving problems of access to in-vehicle data in the
ecosystem of connected and automated mobility.
Section II will ask whether the current discussions
hsurrounding the introduction of data rights or the
use of the data portability right (Art. 20 GDPR) can
offer solutions. This will be followed by an analysis
of whether and how competition law might help
independent service providers to get access to
in-vehicle data (section III). The nal section IV
will suggest that a comprehensive sector-specic
regulatory solution of the governance of in-vehicle
data might be the most promising way for solving
the problems.
82 This problem is also not solved in the U.S.; see for the
U.S. discussion concerning data governance in connected
cars, e.g., Fagnant/Kockelman, Preparing a nation for
autonomous vehicles: opportunities, barriers and policy
recommendations. Transport Research Part A 77, 2015, 167,
178 -180; Anderson et al (n 5) 146; Determann/Perens (n 13)
978-984; Akalu (n 72) 37.
II. Data rights and data portability
42
One group of options for solving data access problems
to in-vehicle data are based upon the possibility of
dening and assigning generally legal rights on
data, which can then also be used for the data of
the connected cars. Due to the many open questions
surrounding the governance of data, broad policy
discussions have emerged about data rights and the
necessity of further legislative initiatives in that
respect. In this section we will focus primarily on
two discussions about possible solutions: (1) The data
portability right of European data protection law,
and (2) the general introduction of new exclusive
and/or access rights on data.
43
According to Art. 20 of the new General Data
Protection Regulation (GDPR), data subjects have a
right to data portability that allows the data subject to
receive their personal data from a data controller in
a structured, commonly used and machine-readable
format, or have them transmitted directly from one
data controller to another. This right should give the
data subjects more control of their personal data,
but also should foster competition between service
providers by lowering switching costs.83 Can this
data portability right be an instrument for solving
the data access problems of independent service
providers in those cases, in which the OEMs have
exclusive control of the in-vehicle data?84 There are
at least three main problems associated with this
solution. A rst general problem of data portability is
that the technical feasibility concerning the meaning
of commonly used formats and interoperability is
so far very unclear. This problem might be solvable
when it comes to data in connected cars, because
standardization regarding in-vehicle data might
be necessary anyhow. A second more difcult
problem is that it is legally very unclear what kinds
of in-vehicle data this right of data portability
would encompass, because most of them are not
uploaded data as in social media but are produced
in the car (often under participation of the OEMs
or component suppliers), or are anonymized data
or business secrets. It is also very doubtful whether
the data portability right would allow for a fast or
even real-time data portability, which would be
83 See Article 29 Data Protection Working Party, Guidelines
on the right to data portability (13 December 2016; rev. on
5 April 2017), 1. For the data portability right as a possible
solution for competition problems caused by exclusive
control of data see Schweitzer/Haucap/Kerber/Welker
(n 59) 183, and, more generally, Graef/Husovec/Purtova,
Data Portability and Data Control: Lessons for an Emerging
Concept in EU Law (TILEC Discussion Paper, 2017-041.
84 Martens/Mueller-Langer (n 7) 25, see the data portability
right as one of the main options for solving the data access
problem to in-vehicle data; for a more general discussion
in regard to the Internet of Things see Urquhart/Sailaja/
McAuley, Realising the right to data portability for the
domestic Internet of Things, Personal and Ubiquitous
Computing, 2017.
Data Governance in Connected Cars
2018
327
3
important for many of the new services in terms of
connected driving.
85
A third important problem is
that this solution might lead to very high transaction
costs, both for the car owners for exercising their
right as well as for the independent service providers
for convincing a sufciently large number of car
owners to use this right for making market entry
protable.86 Therefore the new data portability
right of the GDPR is theoretically a very interesting
option for solving competition problems due to a
lack of access to in-vehicle data, but there are still
too many open technical and legal problems for
making this solution workable in the next years.
It might presumably also require sophisticated
regulatory solutions for lowering the transaction
costs sufciently.87
44 Does the recent general discussion concerning the
introduction of a new property-like right on machine-
generated data or new mandatory access rights to
data offer a solution for the access problems to in-
vehicle data? The intensive discussion surrounding
a new IP-like exclusive right on machine-generated
data with the ensuing proposal of the EU Commission
of a “data producer right” that should be assigned
to the owner or user of a smart device has led to a
broad consensus that the introduction of such an
exclusive right cannot be recommended.
88
After a
consultation the EU Commission has also decided
not to pursue this proposal of such a general “data
producer right”. In the same way, the proposal of a
general mandatory access right to privately held data
(under FRAND conditions) was much criticized and
abandoned by the Commission, although the basic
idea of facilitating more access and reuse of data
has been broadly welcomed both in the academic
discussion and by stakeholders in the consultation.
One important result of this discussion is reluctance
surrounding mandatory solutions compared to much
more favored voluntary solutions for facilitating
85 For a discussion of legal problems of data portability of in-
vehicle data, see Störing, What EU legislation says about car
data, Legal Memorandum on connected vehicles and data,
2017.
86 See Schweitzer/Haucap/Kerber/Welker (n 59) 183.
87 In the telecommunication sector the portability of
phone numbers is facilitated through specic rules in
telecommunication regulation.
88 For this discussion, see Zech (n 36), Drexl, Designing
competitive markets for industrial data: Between
propertisation and access, Max Planck Institute for
Innovation and Competition Research Paper No. 16-13, 2016,
Wiebe, Protection of industrial data - a new property right
for the digital economy? GRURInt, 2016, 877-884 from a legal
perspective, and Kerber, A new (intellectual) property right
for non-personal data? An economic analysis. GRURInt, 2016,
989-998 from an economic perspective; for in-vehicle data
see Hornung/Goeble (n 35), and more general for mobility
data BMVI, Eigentumsordnung für Mobilitätsdaten? Eine
Studie aus technischer, ökonomischer und rechtlicher
Perspektive, 02.08.2017.
contractual solutions about more access to data.
89
The other important conclusion is that the economic
benets and costs of both exclusive rights and/or
access rights are so different between different
sectors and business models that nding general
solutions for dening and assigning new data rights
seem to be extremely difcult or even impossible.
Therefore a broad opinion has emerged that prefer
more sector-specic tailor-made data governance
solutions (see section IV).90 Therefore, the general
discussion about the introduction of data rights do
not seem to offer a clear perspective for solving the
problems of access to in-vehicle data.91
III. Competition law
45 Since the controversial discussion about the access
to in-vehicle data in the extended vehicle concept
focuses on competition problems on the markets
for aftermarket and complementary services,
competition law seems to be an obvious candidate
for nding a suitable policy solution. It is surprising
that so far competition law solutions for granting
access to data have not played a prominent role in
the policy discussion about in-vehicle data.
92
This
section can only present a brief overview about the
options that competition law might offer.
46 In section C.IV. we have seen that in the extended
vehicle concept, the exclusive (monopolistic) control
of the OEMs about the access to the in-vehicle data
89 For these proposals in the Communication, see “Building
a European data economy”, the ensuing consultation and
discussion EU Commission 2017 (n 4), EU Commission,
Synopsis report. Consultation on the “Building a European
data economy” Initiative, 2017, EU Commission 2018 (n 4),
Drexl, Neue Regeln für die Europäische Datenwirtschaft? Ein
Plädoyer für einen wettbewerbspolitischen Ansatz, NZKart,
2017, 339 (Part 1) and 415 (Part 2), Kerber (n 4), Schweitzer/
Peitz. Ein neuer Ordnungsrahmen für Datenmärkte? Neue
Juristische Wochenschrift, 2018, 275-280, and Specht/
Kerber (n 2) 69-99,151-169.
90 See Drexl (n 89) 415, 419, and Kerber (n 4), 109, 133.
91 It will be interesting to see whether the emerging discussion
concerning mandatory access to large anonymized data
sets for training algorithms in the context of articial
intelligence (AI) and machine-learning will lead to new
legislative efforts for introducing general access rights for
these purposes. For a proposal of mandatory data-sharing,
see Mayer-Schönberger/Ramge, Reinventing Capitalism in
the Age of Big Data, 2018, 166-171.
92 However, the results of the consultation about the
Communication “Building a European data economy” have
shown that many rms who have problems with regards to
access to data are skeptical about the extent that competition
law can help to solve data access problems, especially
for small rms in situations with “unequal bargaining
power”. The results of the consultation suggest that this
kind of problem emerges especially in the automotive
sector. See EU Commission, Annex to the synopsis report.
Detailed analysis on the public online consultation result on
“Building a European data economy”, 2017, 13.
2018
Wolfgang Kerber
328
3
(and/or the car) can foreclose competition on the
markets for those aftermarket and complementary
services for which this access is necessary. If –
as our preliminary analysis suggests – systems
competition between entire bundles of connected
cars and services does not work sufciently, then
no undistorted competition on these markets for
aftermarket and complementary services can be
expected, and an obligation of the OEMs for granting
access to the in-vehicle data (e.g., under FRAND-
conditions) might be an appropriate remedy from
a competition economics perspective. The existing
sector-specic obligation of OEMs for granting non-
discriminatory access to repair and maintenance
information in the type approval regulation is
already such a solution (see section B). It can be asked
whether this solution of mandatory access rights to
in-vehicle data for independent service providers
can also be achieved by applying the general rules
of competition law in order to protect competition
on markets for all aftermarket and complementary
services. Although so far no competition law cases
exist concerning obligations to grant access to data,
the increasing interest in the role of data in the
digital economy has led to new discussions about
solutions for data access problems in competition
law. In a recent study about “Modernizing the law
on abuse of market power”, the author (jointly with
Heike Schweitzer, Justus Haucap, and Robert Welker)
analyzed to what extent current European and
German competition law might lead to obligations for
granting access to data in digital contexts, especially
also in IoT-applications (as the connected car).
93
The
following paragraphs try to apply the results of this
study to the problem of access to in-vehicle data.
47
Can the refusal of an OEM to grant access to
exclusively held in-vehicle data be an abusive
behavior according to Art. 102 TFEU by applying
the essential facility doctrine? Whereas there is
a well-established case group of applying Art.
102 TFEU to refusals to grant access to physical
essential facilities (as infrastructure) and to license
IP rights, the essential facility doctrine has so far
not been applied to the refusal to grant access to
“essential” data sets.94 Usually the requirements
for applying the essential facility doctrine are very
high. However due to the economic characteristics
of data, especially non-rivalry in use and the fact
that the incentives for data production are often
much less important than in the case of physical
infrastructure and innovations, the essential facility
doctrine can be applied much more exibly when it
comes to data.95 Besides the requirement of market
93 See Schweitzer/Haucap/Kerber/Welker (n 59) 158-191.
94 See, e.g., Autorité de la Concurrence / Bundeskartellamt,
Competition Law and Data, 2017, 18; Schweitzer/Peitz (n 89)
81; Drexl (n 88) 46.
95 See Schweitzer/Haucap/Kerber/Welker (n 59) 171.
dominance of the data holder, the data have to be
indispensable for offering the service, and the refusal
has to lead to a threatening of the elimination of
competition. If we assume that the relevant market
are the brand-specic markets for aftermarket and
complementary services (i.e. no system markets
exist), the OEMs can be seen as dominant rms,
and their exclusive control of the in-vehicle data
can eliminate competition on these markets.96 The
additional criterion of a “new product” might not
be a problem because of the new innovative services
that are expected to be offered by the independent
service providers. The last criterion is whether the
OEMs have a justication for the refusal of access. We
have seen that safety and cybersecurity concerns do
not provide such a justication. More difcult is the
question concerning the incentives for producing
the data and covering the operating costs of the
entire communication infrastructure. Since the car
users are also participating in generating the data
and have paid for the car and for additional services
(of the OEMs), it is not clear whether and to what
extent such an obligation would lower the incentives
for data production. In addition to that, OEMs can
also be compensated for their (operating) costs.
Much more important is that the consent of the car
users is often necessary for complying with EU data
protection law. Overall, it can be concluded that it
might be possible that the refusal of OEMs to grant
access to in-vehicle data to other stakeholders in the
ecosystem of connected driving can be an abusive
behavior according to Art. 102 TFEU.97
48 However, since the requirements for the “essential
facility” doctrine regarding data are still high
(despite the possibility of more exibility), the
question arises whether there are other options in
competition law. In the above-mentioned study we
particularly analyzed whether § 20 (1) GWB of the
German competition law can also be used for claiming
access to data. § 20 (1) GWB extends the prohibition
of abusive behavior of dominant rms in German
competition law also to rms with so-called “relative
market power”; namely, rms from which other
small or medium-sized rms are dependent, because
they have not sufcient and reasonable possibilities
of switching to other rms. This provision of German
competition law has been used for a long time in
order to solve specic market power problems
below the threshold of market dominance. One of
the case groups are rms (as authorized dealers)
96 For an overview about court decisions in regard to market
dominance of OEMs in aftermarkets and the reluctance of
courts to accept system markets in the automotive industry,
see Schweitzer/Haucap/Kerber/Welker (n 59) 167-180.
97 It is also possible to ask whether the exclusive control of the
OEMs to the connected car, which impedes interoperability
(“closed” car), might be under certain conditions an abusive
behavior of a dominant rm. For such an “interoperability
obstruction”, which also increases lock-in problems, see
Kerber/Schweitzer (n 65) 55.
Data Governance in Connected Cars
2018
329
3
that have specically invested into the relationship
with another rm, and therefore have become
dependent on this rm (“unternehmensbedingte
Abhängigkeit”).
98
Can the refusal of OEMs to grant
access to in-vehicle data to independent providers
of aftermarket and complementary services also be
an infringement of § 20 (1) GWB? Whilst no cases
regarding access to data exist so far, it can be argued
that under certain conditions rms on aftermarkets
and in IoT-contexts with several stakeholders that
need access to the same data for offering valuable
services might claim access to the data that one
stakeholder holds exclusively. In that respect a new
case group relating to access to data in value creation
networks (as in connected cars) might be possible.
The advantage of using this provision is that the data
holder need not be deemed as dominant according to
Art. 102 TFEU or § 18 GWB (in German competition
law). However, it will require much more research in
order to clarify the specic conditions under which
such an obligation for granting access according to §
20 (1) GWB can be justied.99 Therefore, in Germany
§ 20 (1) GWB might offer another way for solving
data access problems in the ecosystem of connected
driving.
49 Therefore, competition law might offer interesting
options for solving problems of access to in-vehicle
data in those cases, in which the OEMs have exclusive
control of these data, e.g. through the application of
the extended vehicle concept. However, these case
groups still have to be developed and it will need
time to clarify the criteria that have to be taken into
account for the necessary balancing of the potential
positive and negative effects of mandatory data
access rights that are based upon either European
or German competition law provisions against
abusive behavior of rms with market power.
Another serious problem is that it might be difcult
and expensive, especially for small- and medium-
sized companies, to enforce access to in-vehicle
data in private litigation. Although more public
enforcement through competition authorities
would be helpful, the instrument of ex-post control
of abusive behavior of powerful rms is always a
difcult and lengthy process for solving problems.
Therefore, it can be asked whether competition law
can also provide instruments outside the control of
abusive behavior. One approach might be the use
98 For this provision in German competition law and
its application, see Nothdurft, Relative Marktmacht:
Gutachten zu Grundlagen, Bedeutung, Wirkung und Praxis
der deutschen Missbrauchsverbote gegenüber relativ
marktmächtigen Unternehmen, 2015, available at
www.faire-importpreise.ch/pdf/gutachten.pdf>.
99 For a deeper discussion, see Schweitzer/Haucap/Kerber/
Welker (n 59) 172-191; due to a possible gap and for
clarication we have made a proposal for amending § 20
(1) GWB of the German competition law for facilitating
data access solutions, especially in Internet of Things
constellations (as also the connected car). See ibid. 191.
of the instrument of a block exemption regulation
according to Art. 101 (3) TFEU, in which problems
of data access, such as complex multi-stakeholder
situations of IoT applications, might be addressed,
either more generally or in a more sector-specic
way.100 It can also be asked whether competition law
could directly challenge the exclusive control of data
by the OEMs in the extended vehicle concept. If the
application of the extended vehicle concept by the
OEMs can be seen as a horizontal agreement between
the OEMs about technological and data governance
solutions after an investigation (as discussed in
section C.VII.), the question of the fulllment of the
criteria for exempting this horizontal agreement
according to Art. 101 (3) TFEU will arise. As part of
such an assessment, the competition authorities
could ask about the efciency effects of such an
agreement and whether the exclusive control of in-
vehicle data through OEMs with its negative effects
on competition is necessary for achieving these
benets. The results of our analysis might raise
serious doubts whether the exclusive control of in-
vehicle data can be justied in such an assessment.
IV. Sector-specific
regulatory solution
50 The last two sections have shown that the already
existing data portability right, as well as competition
law, might help to nd solutions for data access
problems that arise through the exclusive control
of in-vehicle data by the OEMs. However, all of
these policy options are still more theoretical ideas,
which so far have not been tried out and which
will need much more research, effort and time
for implementation. Even if the instruments data
portability and granting the right to access data as
remedy against abusive behavior in competition
law can be applied in the ecosystem of connected
driving, it is not clear whether these options can
be used broadly and fast enough for safeguarding
competition on markets for aftermarket and
complementary services. In addition to that, these
policy instruments cannot help much when it
comes to market failures concerning technological
solutions and information and privacy problems
of consumers (sections C.VI. and C.VII.). Although
there is an option to try to solve the different market
failure problems through applications of remedies
from different legal elds such as competition
law, data protection law, consumer law etc., the
complexity of the technological and data governance
problems in this ecosystem is so large that it is very
unclear whether this leads to a satisfactory solution.
Therefore, it might be more promising to try to
100 Block exemption regulations also have the advantage of
allowing the publication of more specic guidelines that
can deal with different kinds of problems.
2018
Wolfgang Kerber
330
3
develop a tailor-made sector-specic regulatory data
governance solution.
51
It can be suggested that the following problems
should be addressed in a sector-specic regulatory
framework:
(1)
Technological framework: Due to the huge
impact of technological decisions on the
question of who has de facto control of data
and can decide on (the conditions of) their use, a
regulatory framework should encompass policies
for promoting technologies that support a better
use of data, less competition problems, and also
more privacy-friendly solutions regarding the
protection of personal data. In that respect,
the development of solutions for interoperable
on-board application platforms might be
particularly important. These technological
solutions should be seen as part of the long-term
development of the over-arching technological
architecture of connected, automated and
later autonomous mobility. This will require
far-reaching solutions when it comes to
interoperability and standardization (especially
also concerning safety and cybersecurity
problems). Due to the ongoing and technological
evolution, a sophisticated strategy is necessary
for enabling the benets of interoperability and
standardization without impeding innovation.
(2)
Data access: Depending on the developing
technological solutions, specic regulatory
solutions about the governance of the in-
vehicle data might be appropriate. As long
as external server solutions for the in-
vehicle data are applied, regulatory solutions
regarding the access to these data might be
necessary for solving competition problems on
market for aftermarkets and complementary
services. One option can be a broadening of
the current regulated access solution for repair
and maintenance information to all service
providers that need in-vehicle data in the
ecosystem of connected driving. Another option
is the already much discussed “shared server”
solution, which would put all the in-vehicle
data under the control of a neutral entity with
the idea of granting non-discriminatory access.
The question concerning the institutional
design of such a “shared server” also opens up
the discussion about larger data pool solutions
that can also be linked to new ideas of data
trustee solutions. Another solution might be
sector-specic regulations for making the data
portability right an effective instrument for
solving data access problems.101 Also sector-
101 See, e.g., the sector-specic solution in the second Payment
Services Directive (PSD2), through which third-party
payment service providers with the consent of the account
specic rules about access to certain kinds of
in-vehicle data for public authorities (trafc
regulation, law enforcement etc.) might be part
of these data access rules.
(3)
Data economy and privacy: Different
technological solutions such as the on-board
application platform would also enable different
kinds of markets for data, since access to data
could be obtained directly from the car users
leading to new platforms for trading data.
Therefore, the regulatory framework for in-
vehicle data could support the emergence of
these trading platforms. However, even if the
privileged position of the OEMs is eliminated,
complex problems related to dealing with
different types of data have to be solved. This
refers rst and foremost to personal data and
the protection of the privacy of car users, where
the aforementioned market failure problem
might lead to the need of regulatory solutions
for contracts regarding the provision of data
and a minimum of privacy options for car users.
But also, sector-specic rules about data that
can be deemed as business secrets might be
helpful. A sector-specic approach would also
allow regulatory solutions for exploiting the
advantages of data aggregation; specically, that
data analytics and AI can get access to a large
pool of in-vehicle data to increase the quality
of the results (e.g., relating to trafc safety) or
for a better training of algorithms.
52
The advantage of a sector-specic regulatory
framework is that all of these questions are
interrelated with each other, and that therefore the
complex trade-offs between the costs and benets of
different solutions for the governance of these data
might be solved better in an integrated approach.
E. Perspectives
53
The discussion surrounding access to in-vehicle data
and resources is a very important policy discussion,
because it raises many questions that are relevant
in other areas of the digital economy, and especially
in the future world of the “Internet of Things”, in
which the production of sensor data will be nearly
ubiquitous in the ofine world. Smart manufacturing
and smart retailing, smart home, and smart cities are
some of the most important examples in that respect.
In all of these areas it is so far very unclear how an
appropriate data governance framework should look
like. However, in all of these contexts very similar
questions will arise as they have been discussed
owners might get access to bank account data for offering
their services to the consumers.
Data Governance in Connected Cars
2018
331
3
here with regards to the data in the ecosystem of
connected and automated cars.
Acknowledgements
I thank the participants of the Ascola conference on
June 22, 2018 (NYU Law School, New York) and the EPIP
conference on September 6, 2018 (ESMT, Berlin) as well as
Daniel Möller for valuable comments. The research for this
paper has been funded by University of Marburg leading
to no conict of interest. Other research for studies related
to this topic have been funded by Bundesministerium
für Bildung und Forschung [Specht/Kerber: Datenrechte
– Eine rechts- und sozialwissenschaftliche Analyse
im Vergleich Deutschland – USA, 2018, as part of the
ABIDA (Assessing Big Data) project, University Münster]
and Bundesministerium für Wirtschaft und Energie
[Schweitzer/Haucap/Kerber/Welker: Modernisierung
der Missbrauchsaufsicht für marktmächtige
Unternehmen, 2018].

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT