See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321060200

Standardization for the Digital Economy: The Issue of Interoperability and

Access Under Competition Law

Article in The Antitrust Bulletin · December 2017

DOI: 10.1177/0003603X17733359

CITATIONS READS

3 456

1 author:

Björn Lundqvist
Stockholm University
26 PUBLICATIONS 62 CITATIONS

SEE PROFILE

All content following this page was uploaded by Björn Lundqvist on 31 January 2019.

The user has requested enhancement of the downloaded file.

      1  

Standardization  for  the  Digital  Economy  

 
-­‐  The  Issue  of  Interoperability  and  Access  under  Competition  Law  
 
By  
 
Björn  Lundqvist  
 
Abstract    
 
This article discusses several aspects of the Digital Economy. Firstly, the Intellectual
Property landscape of the soon to come Internet of Things is drawn up, discussed
and scrutinized, and the current issue whether non-personal digital data should
become a property right, in-itself, is analysed. Secondly, in light of the IP landscape,
“pictured” in the first part of the article, the current standardisation efforts for the
Digital Economy are discussed, e.g. what are the challenges, and how much should
be standardized and how lenient should competition authorities treat pre-standard
consortia. Thirdly, current and future competition law issues for the Digital Economy
are identified, and lastly the unworkable dichotomy between personal data, non-
personal data is criticized. The paper conclude that general competition law may not
be readily available for accessing generic (personal or non-personal) Data, except for
the situation where the Data set is indispensable to access an industry or a relevant
market; while sector specific regulations seem to emerge as a tool for accessing
Data held by competitors, third parties and possibly competing ecosystems.
However, the main issue under general competition law in the Data industry, at its
current stage of development, is to create a levelled playing field by trying to facilitate
the implementation of Internet of Things; thus, competition authorities should be
cautious about the current consortia driven standard-setting movement in the Digital
Economy, when the technology being standardized is not infrastructure type, but
rather of upper layer substitute data interoperability technical solutions.

Keywords: Internet of Things, Industrial Internet, Standardization, SEP, Competition

Law, Antitrust, Big Data, Open Data, Intellectual Property Law, Privacy, Data
Protection, PSI, Platforms, Ecosystem
 
 
1.  Introduction    
 
The Digital economy, the ‘Internet of Things’ scenario and Big Data are currently
triggering the interest of politicians, businessmen, the academic community and,
even, the general public. The groups are interested for different reasons; for
example, businessmen see an opportunity for the creation of wealth, researchers see
the possibility of discussing issues of “paradigm shift”, globalisation, and “change”,
while everyone acknowledge that the collection and distribution of personal Data may
raise both privacy and data protection concerns. The stakes are also high. According

Electronic copy available at: https://ssrn.com/abstract=2977242

      2  

to the EU Commission in 2014 alone, cross-border data flow generated estimated 2.8
trillion in economic value, exceeding the value of global trade in goods.1 Cross-
boarder data flows are the fastest growing component of EU as well as US trade, and
a study found that inbetween 2008 to 2012 the cross-border flow of trade increased
with 49 %, while the increase of cross-border trade in goods was only 2.4 %.2

The interface between the legal systems triggered by the creation, distribution and
consumption of Data in the digital economy is difficult to grasp, and this paper
therefore tries to dissect this interface. The paper starts with the attempt to identify
what legal systems are applicable in this process. What intellectual property law
system may be applicable when Data is obtained from devices, sent to other devices
and/or distributed to the Cloud, and, ultimately, when it is re-used. Who ‘owns’ data,
and do data protection rules create nascent property rules regarding personal data?
Secondly, the paper will discuss when where and what should be standardized.3
Thirdly, the paper specifically focuses on the application of competition law vis-a-vis
the firms included in the standardization of the digital economy. Should consortia
driven standard-setting be promoted or benefit from heighten antitrust scrutiny. In
addition to this discussion, the other issue regarding standardization is discussed,
access. May competition law be used to gain access to Open or Big Data, or the
infrastructure around that Data?4 May competition law be used to create a level
playing field between holders and non-holders of essential data?

The paper conclude that general competition law may not be readily applicable to
access generic data, except for the situation where the data set is indispensable to
access an industry or a relevant market;5 while sector specific regulations seem to
emerge as a tool for accessing Data held by competitors or firms in general.
                                                                                                               
1  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  
2  Ibid.  
3  See  Ezrachi  and  Stucke,  Virtual  Competition  The  Promise  and  Perils  of  the  Algorithm-­‐­‐Driven  

Economy,  (Harvard  University  Press,  2016);  and  Stucke  and  Grunes,  Big  Data  and  Competition  Policy,  
(Oxford  University  Press,  2016).  
4  In  reference  to  Open  Data  see  Drexl,  “The  Competition  Dimension  of  the  European  Regulation  of  Public  

Sector  Information  and  the  Concept  of  an  Undertaking”,  in:  Drexl  and  Bagnoli  (eds.),  State-­‐Initiated  
Restraints  of  Competition  (ASCOLA  competition  law),  (Edward  Elgar,  Cheltenham,  UK;  Northampton,  MA  
2015),  64  -­‐100.  See  also  Lundqvist,  “”Turning  Government  Data  into  Gold":  The  Interface  Between  EU  
Competition  Law  and  the  Public  Sector  Information  Directive  -­‐With  Some  Comments  on  the  Compass  
Case”  in  44  I  I  C  -­‐International  Review  of  Intellectual  Property  and  Competition  Law,  (Nr.  1,  2013),  79-­‐­‐
95.  
5  ECJ,  Huawei  Technologies,  C-­‐170/13,  ECLI:EU:C:2015:477.  

Electronic copy available at: https://ssrn.com/abstract=2977242

      3  

However, the main issue under competition law in the Data industry, at its current
development, is to create a levelled playing field by trying to facilitate the
implementation of Internet of Things. For example, we might see brick-and-mortar
firms develop into collecting, storing and trading Data in competition with the
incumbent e-platform firms. Such a development could increase competition in the
Data industry to the benefit of consumers and society. However, if we can predict
that this might happen, what consequences might that have for the use of
competition law vis-à-vis the incumbent firms in the Data industry, hence those firms,
e.g. Google, that are under investigation for being dominant.

2. Background

The creation, collection, storing of, commercially and technically using, and the
dissemination of Data, be it government (open) data and/or private (big) data 6 ,
require a number of components to materialise. Most importantly, the data industry to
materialize needs someone that wants to invest in the collection and storing of Data.
Until lately, a government authority or a similar body has usually been the data
collector. Hence, a Public Sector Body, based on an obligation in law, collecting for
society necessary data concerning, for example, land ownership, trademarks,
weather information, maps or company data, storing the Data on servers, has up until
recently been the ‘normal’ data collector. While these public collectors have
gradually, voluntarily or under the PSI Directive, started to market the Data to
consumers on the Internet and to firms reusing the same (e.g. Data brokers, or Data
re-users); private parties, except a few, have not had the interest nor the means to
collect and store vast amount of Data in a similar way.7

However, with Google and Facebook, possibly, being the pioneers, private entities
are clearly starting to collect and store large volumes data and then mostly personal

                                                                                                               
6  The  definition  of  Big  Data  is  vague  and  lacks  precision,  see  De  Mauro,  Greco  and  Grimaldi,  (2016),  122-­‐

135.    
7  For  general  information  see  OECD,  ‘Data-­‐Driven  Innovation-­‐Big  Data  for  Growth  and  Well-­‐Being’,  (6  

October,  2015);  For  the  costs  of  the  necessary  investments  to  access  the  data  industry  see  OECD,  ‘Big  
Data–Bringing  Competition  Policy  to  the  Digital  Era’,  (30  September  2016).  
      4  

(consumer) data.8 Private parties, active online selling adds, or, more accurately,
access to avenues for the marketing of goods or services to potential consumers,
have realised that it may be profitable to collect personal data. Mainly consumer
data, but also other forms of information and knowledge are collected. These firms
want to be able to sell focused add spots or avenues that, with the use of the
collected user data and algorithms, pin-point the most likely purchasers and also
warrants that the commercial message penetrate and reach the intended focus
groups.9

In addition, some firms are today actively seeking out unique Data sets, especially in
the cultural arena, to gain more traffic in their own ecosystems, so collect even more
personal data and thereby boost their marketing service vis-à-vis firms wanting to
market on-line.10 Google has, for example, been accused of leveraging their potential
market power based on holding vast amounts of data by expanding its business,
from general search, into new services/products, provided to users at zero price,
while the goal of these new services/products is to obtain more data from users.11
Indeed, Google is active on two levels or markets, downstream providing diversified
on-line services to consumers, while also being active upstream on the Add market,
where the Data collected from the zero price services is used to obtain a competitive
edge.

It is clear that, especially Google, but also other e-platforms and ecosystems (for
example, Apple, Amazon, Microsoft, Facebook and to some extent Spotify) are in the
business of collecting personal data which is either voluntarily provided by users or
collected from users based on their conduct, e.g. general and special searches,
cookies etc. Moreover, several economists and lawyers have suggested that the

                                                                                                               
8  Google  is  involved  in  several  investigations  in  several  jurisdictions  regarding  the  company’s  business  

conduct.  For  example  the  Commission  Statement  of  Objections  to  Google  on  Android  operating  system  and  
applications,  20  April  2016.  Cf.  http://europa.eu/rapid/press-­‐release_IP-­‐16-­‐1492_en.htm  
9  Ibid.    
10  See  the  updated  PSI  Directive  now  including  museums,  libraries  and  other  cultural  institutions,  and,  

moreover,  the  awkward  exemption  for  exclusive  license  for  cultural  data  bases.  Cf.  The  Directive  on  the  
re-­‐use  of  public  sector  information  (Directive  2003/98/EC,  known  as  the  'PSI  Directive')  entered  into  
force  on  31  December  2003.  It  was  revised  by  Directive  2013/37/EU  which  entered  into  force  on  17  July  
2013.  
11  Newman,  ‘Search,  Antitrust  and  the  Economics  of  the  Control  of  User  Data’  30  Yale  Journal  on  

Regulation,  (No.  3,  2014),  3  et  seq.  with  references.  Available  at  SSRN:  http://ssrn.com/abstract=2309547  
or  http://dx.doi.org/10.2139/ssrn.2309547.  (last  visited  14  December  2016).  
      5  

collection of personal data causes these firms to gain and hold market power.12
Indeed, that the market power obtained by holding vast amount of Data create
insurable barriers to entry for second movers and even that the market for providing
services based on data may tip due to network effects based on the holding of vast
amount of Data.13 The vast amount of data collected increases the quality of the
service, which in turn generates more users to the service. In that connection
especially Google is under investigation for abusing their market power by several
competition authorities.14

While, these first movers in the Data industry may hold market power and are
intensively competing so to position themselves for the upcoming IoT paradigm, it is
clear that the brick-and-mortar industries will start monitor users and collect Data
when everything will become connected. The brick-and-mortar firms may then try to
enter the data industry based on their market position for the specific devices they
produce, or they may buy in to certain existing digital ecosystems that provide them
with necessary interfaces (software), Cloud space, algorithms (also software). 15
Indeed, brick-and-mortar firms would either go on their own and create their own
ecosystems (based on their own technical standards) or become vertically integrated
into pre-existing networks by making their devices compatible with certain technical
solutions provided by for example Google, Microsoft or Amazon.

What will happen when car manufacturers, refrigerator producers, etc., start
collecting data based on their products is difficult to predict, but it may cause
something similar to ’multi-homing’. Indeed, that general or generic consumer data
regarding use, data traffic, consumption, gps position etc., will be collected by several
firms. This in turn could lead to increase sale/license of Data or database
connections for re-use of data, similar to what has happened in the PSI sector. Thus,
that markets for collection, storing and selling/licensing Big Database access will
                                                                                                               
12  Ibid.  See  also  e.g.  Kerber,  ‘Digital  Markets,  Data,  and  Privacy:  Competition  Law,  Consumer  Law,  and  Data  

Protection’,  MAGKS,  Joint  Discussion  Paper  Series  in  Economics,  (No.  14-­‐2016).  Available  at    
SSRN:http://ssrn.com/abstract=2770479  or  http://dx.doi.org/10.2139/ssrn.2770479  (last  visited  12  
December  2016)  3  et  seq.    
13  Ibid.  See  also  OECD  (2016);  The  German  and  French  Competition  Authorities  (2016),  7,  10  et  seq.  
14  For  example,  the  Commission,  Statement  of  Objections  to  Google  on  Android  operating  system  and  

applications,  20  April  2016.  http://europa.eu/rapid/press-­‐release_IP-­‐16-­‐1492_en.htm    

15  OECD  (2016)  op  cit,  supra  note  7.  The  German  and  French  Competition  Authorities  joint  paper,  

“Competition  Law  and  Data”,  (10  May,  2016),  10  et  seq.    
      6  

evolve and that large brick-and-mortar firms will enter this market in competition with
the incumbent Google, Microsoft; Amazon etc.16

For brick-and-mortar firms to start collecting data, the infrastructure needs to be in

place, at least if we want private parties to work themselves up vertically from
production of devices and also become the creators and collectors of data sets. Here
we see the revolutionary aspect of what we today call ‘the Internet of Things’ (IoT).
IoT essentially comprises of client side devices with different sensors (speed,
acceleration, motion, proximity, location, distance, weight, humidity, altitude, depth,
compass, temperature, pressure, still image, video image, infrared, audio, noise
level, blood sugar, heart rate, number of steps, etc.) that are connected to a server
side through the Internet. Of course, technically, this is not something entirely new.
Previously, the IoT technology was marketed under the concept of M2M (machine to
machine). Both devices with sensors and the telecom networks could handle M2M,
while the M2M did not succeed because of ‘back-office’ technology problems.17 ‘The
Cloud’, i.e. the computing, servers and the client server interface, were not up for the
challenge. The capacity to, very quickly and inexpensively, store massive volumes of
data was not available. This have changed in the up-coming IoT era since computer
firms have in the last 10-15 years invested in R&D in the server-client technology and
the client server technology is now mature enough to be implemented.18 So, what
firms will control the client server interface? According to some studies, the patent
distribution in the IoT domain is very fragmented with the top patent filer in the field
holding around 5% of the total patents.19 Accordingly, LG holds the largest patent
portfolio and, it is closely followed by Ericsson and Qualcomm. Others claim that
Microsoft holds the largest portfoilio in reference to the client server interface. 20
Possible, there is, thus, some confusion about who’s technology will be used for the
IoT standards. Nonetheless, the client server interface technology will be protected
                                                                                                               
16  Regarding  the  potential  benefits  of  trading  Data  see  Lundqvist,””Turning  Government  Data  into  Gold":  

The  Interface  Between  EU  Competition  Law  and  the  Public  Sector  Information  Directive-­‐With  Some  
Comments  on  the  Compass  Case”  in  44  I  I  C  -­‐International  Review  of  Intellectual  Property  and  
Competition  Law,  (Nr.  1,  2013),  79-­‐95,  79-­‐81.    
17  Östman,  “The  IP  of  Things”,  (2016)  published  on  LinkedIn  https://www.linkedin.com/pulse/ip-­‐­‐

things-­‐niklas-­‐%C3%B6stman?trk=prof-­‐post.    (last  visited  18  September  2016).      

18  Ibid.      
19  Lexinnova,  Internet  of  Things,  Patent  Landscape  Analysis  (2014),    

http://www.wipo.int/export/sites/www/patentscope/en/programs/patent_landscapes/documents/inte
rnet_of_thing  s.pdf  (last  visited  25  November  2016),  3  et  seq.  
20  Östman,  (2016),  1  et  seq.  
      7  

by patents and, moreover, the upcoming 5G and the current 4G, i.e. the network
infrastructure technology, could even be considered suffering from early signs of a
patent thicket fatigue.21 The 5G telecom network that is currently being rolled out
seem to revolve around technology from the incumbent firms, e.g. Ericsson, Nokia,
while also including technology originating from Huawei, and they hold large patent
portfolios including several SEP 5G patents.

Thus, the Cloud, the alogirthms and the interface between devices, 5G telecom and
the Cloud are the new ‘things’ in the Internet of Things. A new ‘thing’ is also the need
for interoperability and technical standards in reference to interoperability and these
interfaces. It is telling to see that CEN and CENELEC are no quickly adopting to the
new paradigm and are implementing IP Guidelines and SEP policy briefs. 22
Moreover, given the waste amount of patents and software needed for the IoT, we
will presumably will see several SEP litigations in the future.

For the brick-and-mortar firms to evolve into data processing firms, they need to
develop their own internet based ecosystems, including machine-learning algorithms
and Clouds. Otherwise, the incumbent firm, such as Amazon, Google and Microsoft,
will provide these services and the brick-and-mortar firms will only provide the hard
ware, i.e. the devices, for obtaining big data.23 The industry would then evolve in the
direction of more collaborations between the incumbent data firms’ respective
ecosystems and the brick-and-mortar firms. Presumably, exclusive or semi-exclusive
vertical arrangements between device producers and certain firms controlling their
respective ecosystem may evolve. Such vertical relationships may under certain
criteria benefit competition but from a business strategic perspective, the brick-and-
mortar firms need to be cautious which ecosystem they decide to get involved with.
We seem to see that data firms, telecom firms and brick-and-mortar firms are
teaming up in consortia so that to develop interoperability pre-competition and pre-

                                                                                                               
21  Ibid.    
22    Cf.  CEN  and  CENELEC  position  paper  on  standard  essential  patents  and  fair,  reasonable  and  non-­‐­‐

discriminatory  (frand)  commitments,  (September  2016),  

http://www.cencenelec.eu/News/Policy_Opinions/PolicyOpinions/EssentialPatents.pdf    (last  visited  18  
September  2016)  
23  OECD  (2016).  The  German  and  French  Competition  Authorities  (2016),  14.  
      8  

standard.24 However, the early signs we see now on high concentration in the Data
industry could then presumably be ascertained depending on whether we would like
competition to thrive in the Data industry.25

Interoperability technologies, like the IoT, promise that devices of everyday life can
communicate with each other and that such communication can be stored in ‘The
Cloud’, i.e. in the increasing global data storing capacity. The large capacity can be
attributed to client server technology and developing examples of increasing M2M or
device-to-device communication (“D2D communication”) are abounded. In the
automotive industry, the eCall machines already today opens up for the possibility of
obtaining all sort of information regarding the car and the driver, while car
manufacturers conduct R&D on apps or driverless systems that require multiple
connected devices to work together (e.g., sensors, radars, high-powered cameras,
etc.).26 Smart-cities, smart buildings, smart-grids and electro mobility converge with
mechanical engineering, logistics and seamless wireless communications to provide
new functionalities for businesses and consumers, including lighting, air quality
control, security and surveillance, traffic management, etc. Thus, data may be
collected as by-product and stored from all forms of devices and communication, and
while the data had a first use, e.g. to guide the car driver or to communicate between
cars, it may be re-used for other purposes.27

In reference to the IoT line of thinking, the ‘device’ industry will be mature industries
inhabited by often large firms that have their own IP portfolios.28 For example, the car
is a ‘device’ in the Data industry, and while the intellectual property rights are

                                                                                                               
24  See  for  example  the  AUDI  AG,  BMW  Group,  Daimler  AG,  Ericsson,  Huawei,  Intel,  Nokia  and  Qualcomm  

Incorporated,  announce    27  September  2016  the  formation  of  the  “5G  Automotive  Association”.  
https://www.ericsson.com/news/160927-­‐telecommunications-­‐and-­‐automotive-­‐players_244039854_c      
25  ”With  companies  such  as  Amazon,  Google  and  Microsoft  providing  machine  learning  algorithms  as  part  

of  their  cloud  computing  services,  small  companies  find  it  increasingly  more  convenient  to  have  their  data  
processed  and  mined  using  external  IT  infrastructures.  Indeed,  Cisco  forecasts  that,  by  2019,  86%  of  all  
business  workload  processing  will  be  processed  by  cloud  computing.  But,  as  a  greater  number  of  
companies  become  dependent  on  the  infrastructures  of  a  few  providers,  the  latter  get  access  to  significant  
volumes  and  variety  of  data  that  allows  them  to  improve  further  their  own  data  analysis  algorithms.  If  the  
trend  continues,  a  competition  problem  may  arise  in  the  future,  as  new  entrants  may  not  be  able  to  build  
sufficiently  powerful  IT  infrastructures  whose  analytical  software  can  compete  with  those  of  incumbents.  
OECD  (2016).  The  German  and  French  Competition  Authorities  (2016),  14.  
26  Cf.  CEN  and  CENELEC  (2016),  1  et  seq.    
27  Ibid.    
28  For  an  explanation  of  patent  thicket,  see  Shapiro  (2001),  119.    
      9  

stacking up in the IoT, the car manufactory industry has likewise densely populated
intellectual property rights landscape.29

Indeed, in light of the above, data, originating from users, from devices, sent through
the 4G or 5G networks to the client servers and the Cloud are heavily boxed in by
intellectual property rights. The intellectual property rights cater to the technologies,
the infrastructure, while the data as such is, probably, not covered by the
infrastructure intellectual property rights. However, firms holding large IP portfolios in
the specific device industry might try to exclude or obtain licensing fees from a new
user trying to access Data flowing in the system or stored in The Cloud connected to
the specific device they produce.30 However, the network providers and the algorithm
providers and, possible also the providers of the specific Cloud (if they are separate
entities) may also try to either access the data collected by the specific device or
even, technically, try to exclude others from gaining access to the Data. Indeed, even
though the data is owned by noone, the firms providing the collection mechanisms, or
other vessels, i.e. the devices or platforms for data, have obtained intellectual
property rights to these. Presumably, in the end, this will probably require these firms
to collaborate.31

3. The Definition of Data

The EU Commission are currently producing several reports and other documents in
relation to the digital economy in general, and, to Big Data, in specific. In these
documents there is a clear dichotomy between, on the one hand, personal data, and,
on the other, non-personal data. The Commission claims that while personal data will
be comprehensively and completely regulated under the General Data Protection
Regulation, non-personal data is rather ‘up for grabs’.32 Indeed, in reference to non-

                                                                                                               
29  Cf.  CEN  and  CENELEC  (2016),  1  et  seq.    
30  There  are  several  news  articles  regarding  for  example  the  collaboration/license  agreements  between  

Microsoft  and  Facebook  regarding  the  Microsoft  Cloud.  Also,  users  of  the  Microsoft  Cloud  needs  to  obtain  
a  license  from  Microsoft,  https://www.microsoft.com/en-­‐us/Licensing/product-­‐licensing/innovations-­‐
for-­‐the-­‐cloud.aspx.  
31  See  for  example  the  AUDI  AG,  BMW  Group,  Daimler  AG,  Ericsson,  Huawei,  Intel,  Nokia  and  Qualcomm  

Incorporated,  announce    27  September  2016  the  formation  of  the  “5G  Automotive  Association”.  
https://www.ericsson.com/news/160927-­‐telecommunications-­‐and-­‐automotive-­‐players_244039854_c      
32  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  

economy,  COM  /2017,  9  Final,  10.1.2017,  11.  

      10  

personal data there is room for trade regulations, competition law and, even, possible
new forms of property rights.

However, what is personal and non-personal data. According to Art 4 of the General
Data Protection Regulation, ‘personal data’ means any information relating to an
identified or identifiable natural person (‘data subject’); an identifiable natural person
is one who can be identified, directly or indirectly, in particular by reference to an
identifier such as a name, an identification number, location data, an online identifier
or to one or more factors specific to the physical, physiological, genetic, mental,
economic, cultural or social identity of that natural person. Non-personal data are all
other form of data that do not trigger this definition.

The definition of personal data is wide since also information that is non-personal
might indirectly in combination with other information identify a natural person and
become personal data. Thus, non-personal, even meta data, can in combination with
other data and with the use of a developed algorithm become personal data under
the definition. It is therefore recommended that firms collecting data, even only meta
data or aggregated data, even in an industrial internet setting, do take the rules
contained in the General Data Protection Regulation in consideration. Indeed, it can
be wise to calibrate the collection mechanism, in for example an industrial internet
setting, so to only transfer and collect general non-personal data and to keep such
data sets intact. And, of course, non-personal data are disconnected from any right
on the behalf of the data subject.

3.1 Noone Owns Data

The data, the information (as such), irrespectively how private and how valuable, is
not currently covered by property right.33 Noone owns personal data, while still the
‘data subject’ in the EU holds some rights to it according to the General Data
Protection Regulation.34

                                                                                                               
33  There  are  authors  that  propose  the  recognition  of  ownership  rights  for  consumers  over  the  data  they  

produce:  Hoofnagle  and  Whittington“Free:  Accounting  for  the  Costs  of  the  Internet’s  Most  Popular  
Price”,    61  UCLA  Law  Review,  (2014),    606-­‐670.      
34  There  some  rights  connected  to  personal  data  in  Articles  18-­‐20  of  the  General  Data  Protection  

regulation,  such  as  right  to  have  data  corrected,  ”right  to  be  forgotten”  and  data  portability.  In  reference  to  
      11  

Notwithstanding this, if individual data (personal or non-personal) fulfils the

requirement for an intellectual property right, e.g. copyright, it can be covered by
copyright (3rd party copyright, or copyright held by firm that is also gate keeper to
accessing the data in the server). Moreover, as discussed above, the firms providing
the ecosystem or infrastructure of the IoT will have the infrastructure covered by
patents and copyright and also, technically, prevent access to the data. Traditionally
copyright owners regularly resort to technical protection measures (TPMs), cf. Art 6
InfoSoc, to prevent access to the copyright protected content.35 Interestingly, InfoSoc
not only find breaching these technical protection measures a copyright infringement,
but also the manufacturing and sale of devices which have the primary purpose or
effect of enabling such circumvention may be a copyright infringement in itself.36

Whether the Data may be covered by rules regarding trade secrets have up until
recently been regulated very differently in the different Member States.37 However,
the regulatory landscape for trade secrets is dramatically changing with the
introduction of harmonised rules based on Directive 2016/943/EU of 8 June 2016 on
the protection of trade secrets. It is probable that Data may be protected under the
rules in the directive. Individual data might not constitute a trade secret, but the
combination of data or information (that as such is not publicly available), data sets,
might well be covered. 38 The same argument also applies with regard to the
requirement of commercial value under the directive.39 Even if the publicly available

                                                                                                                                                                                                                                                                                                                                                         
data  portability,  the  right  is  however  limited  making  it  less  attractive  to  change  social  website  for  
consumers.  Cf.  Regulation  (EU)  2016/679  of  the  European  Parliament  and  of  the  Council  of  27  April  2016  
on  the  protection  of  natural  persons  with  regard  to  the  processing  of  personal  data  and  on  the  free  
movement  of  such  data,  and  repealing  Directive  95/46/EC  (General  Data  Protection  Regulation)  (Text  
with  EEA  relevance)  OJ  L  119,  4.5.2016,  p.  1–88.    
35  See  interesting  conference  paper  by  Ciani,  “A  competition  law  oriented  look  at  the  application  of  data  

protection  and  IP  law  to  the  Internet  of  Things:  towards  a  wider  holistic  approach”,  MPI  Post  Doc  
Conference,  Munich  21  October  2016.    
36  Ibid.  
37  Sweden  is  one  of  few  Member  States  that  have  a  specific  Act  for  the  protection  trade  secrets,  while,  for  

example,  trade  secrets  in  the  UK  and  in  Denmark  have  been  protected  under  case  law  and  the  marketing  
law  (unfair  competition  law),  respectively.  In  Sweden,  collections  of  customer  data,  e.g.  addresses,  have  
been  protected  under  the  Trade  Secret  Act.    
38  See  Drexl,  Hilty,  Desaunettes,  Greiner,  Kim,  Richter,  Surblytė  and  Wiedemann    “Data  ownership  and  

Access  to  data,  Position  Statement  of  the  Max  Planck  Institute  for  Innovation  and  Competition”,  Max  
Planck  Institute  for  Innovation  and  Competition  Research  Paper  No.  16-­‐10,  (2016)  p.  6  et  seq.  
39  Ibid.    
      12  

data as such might not possess commercial value, their combination can acquire a
certain value, conferring on the data holder a competitive advantage.40

Moreover, the interface between the rules of trade secrets and the right to data
portability according to the General Data Protection Regulation is not fully sorted
out.41 While according to the old data protection directive, data subjects could have
an overriding right to transfer personal data, the data portability right according to the
new regulation is less clear on this point.42 The reason for this shift could be that the
General Data Protection Regulation aims to establish a high threshold for data
protection and for the free movement of data, i.e. the fifth freedom of the internal
market.43 Possibly, in reference to this issue, the commercial reasons have been the
overriding goal of data protection.

Database sui generis protection44 may be applicable for holders of Data. Government
authorities and private parties (at least in the Nordic Member States) maintaining
specific data sets in databases, which they commercially provide access to (via the
Internet), e.g. the national trade mark database or the databases for official
addresses, land ownership, weather and maps etc, normally claim database
protection as the basis for requiring re-users or users wishing to access and utilizing
the databases to enter into license agreements.45 Public and Private entities that
collect personal or non-personal data in databases might, thus, also fulfil the
requirements for obtaining database sui generis copyright protection. According to
the Commission, subject to exceptions, use by others (e.g. extraction of the content,
reproduction of re-utilisation of the database) can be prevented by the database
author or maker, but only to the extent that either their database in its entirety or

                                                                                                               
40  Ibid.    
41  Surblyte,“Data  Mobility  at  the  Intersection  of  Data,  Trade  Secret  Protection  and  the  Mobility  of  

Employees  in  the  Digital  Economy,  Max  Planck  Institute  for  Innovation  &  Competition  Research  Paper  No.  
16-­‐03.  (2016),  14  et  seq.  Available  at  SSRN:  http://ssrn.com/abstract=2752989  or    
http://dx.doi.org/10.2139/ssrn.2752989  (last  visited  15  September  2016).    Cf.  Article  20  (4)  and  recital  
63,  General  Data  Protection  Regulation.    
42  Ibid.  
43  The  Swedish  Trade  Council  (2016),  1  et  seq.    
44  Directive  No.  96/9/EC  of  the  European  Parliament  and  of  the  Council,  of  11  March  1996  on  the  legal  

protection  of  databases.    

45  See  ECJ,  Freistaat  Bayern  v.  Verlag  Esterbauer  GmbH,  C-­‐490/14,  ECLI:EU:C:2015:735,  in  reference  to  

maps.,  where  the  CJEU  states  geographical  data  presented  in  maps  can  be  deemed  to  be  "independent  
material"  within  the  meaning  of  Article  1(2)  of  the  Database  Directive,  and  it  enjoys  protection  under  the  
Database  Directive.    
      13  

substantial parts thereof are concerned, cf. article 7(1), or when others seek to use
insubstantial parts of the database in a "repeated and systematic" manner, cf. article
7(5). The Commission concludes that protection offered thus does not apply to
machine-generated data as such, while still be applicable to new and re-used
datasets.46

The Commission rigorously claims that the General Data Protection Regulation
(GDPR) and the lex specialis ePrivacy Directive fully regulate the processing of
personal data. The right of the protection of personal data is a fundamental right in
the EU, according article 8 of the Charter of Fundamental Rights. The GDPR sets
out the rules for processing personal data including i.a. the collection, use of, access
to and portability of personal data as well as the possibilities to transmit or to transfer
personal data. The right to data portability, according to article 20 of the GDPR will,
according to the Commission, allow individuals to obtain copies of personal data they
have provided to a service provider (data controller) and to move that data to another
service provider (data controller), (for example, personal data on a social media). It
may assist avoiding potential lock-in effects.47

In light of the above, it seems clear that the digital economy, the Internt of Things, in
general, and Big Data, in specific, even though not being directly encompassed by a
property right, still are submerged in several layers of intellectual property rights.
Indeed, given the potential risk for ‘thickets’ it might also be concluded that neither
data nor other technical, commercial or intellectual aspects of the developing data
industry need more intellectual rights protection to flourish. 48 The protection is
already on a high level.

                                                                                                               
46  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  

economy,  COM  /2017,  9  Final,  10.1.2017,  20.  

47  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  

economy,  COM  /2017,  9  Final,  10.1.2017,  20.  

48  In  reference  to  the  algorithms  normally  used  to  process  data  in  data  bases.  Drexl  et  al,  op  sit,  state:  

”[t]he  Max  Planck  Institute  for  Innovation  and  Competition  does  not  see  any  need  to  create  special  legal  
protection  of  algorithms  used  in  data  processing  (e.g.  in  the  context  of  big-­‐data  analysis).”  And  continues  
”concrete  computer  programs  for  processing  data  are  already  protected  by  copyright  law  of  the  Member  
States  implementing  Directive  2009/24/EC  of  23  April  2009  on  the  legal  protection  of  computer  
programs.  Nevertheless,  this  protection  covers  neither  the  functionality  of  a  computer  program  (judgment  
SAS  Institute  Inc.,  C-­‐406/10,  ECLI:EU:C:2012:259,  paras  39-­‐41)  nor  the  underlying  general  algorithm  
(which  is  understood  here  as  a  set  of  rules  to  solve  a  problem  step  by  step,  independent  of  its  expression  
and  representation,  e.g.  the  description  of  the  steps  to  be  made  for  analyzing  or  filtering  data  and  the  
criteria  to  be  applied).  This  is  already  implied  by  Recital  11  of  the  Directive,  which  clarifies  that  copyright  
      14  

However, as hinted to above, if noone holds property right to the data there is no
clear rule whether the device producer, the software/hardware producer of the
collecting mechanism, the network providers or algorithm providers or the Cloud
service providers may exclude each other from the data. Since the data flow through
their networks and clouds, can the firms that control these networks and clouds tap in
to the data without violating any firm’s property right? Or, more importantly, who
should trade with who? This might cause competition policy problems. Competition
could possibly be enhanced if the device producers could enter the data industry as
independent stand-alone competitors in an IoT setting. Perhaps, this requires that
they should control (‘own’) the data created by their devices. Possibly, the sui generis
database right would suffice in this regard, at least in reference to whole databases
or data sets, so no new data right need to be created. Nonetheless, the Commission
in one its latest report purport that, in reference to non-personal data, the digital
economy might benefit from a property right given to the data producer. 49 The
Commission continues and discuss whether a property right or at least a defensive
right in the data should be created. 50 In reference to database protection, the
Commission state that there is a lack of legal certainty entailing the absence of legal
protection in relation to non-personal or anonymised machine-generated data not yet
structured in a protected database.51 It seems that the Commission, hence, is aiming
for data that have not yet been included in a database, data which is, by the device
producer, the software/hardware producer of the collecting mechanism, the network
providers or algorithm providers or the Cloud provider, ”picked up” on the way to the
database, or when the database is not yet a database according to the sui generis
database regulation. Whether this is a real concern still remains unclear. However,

                                                                                                                                                                                                                                                                                                                                                         
protection  for  computer  programs  should  not  extend  to  the  “ideas  and  principles  which  underlie  any  
element  of  a  program”.  Some  economist  have  suggested  a  property  solution,  cf  Hoofnagle  and  Whittington  
(2014),  606-­‐670.      
49  Commission,  Communication  from  the  Commission  to  the  European  Parliament,  the  Council,  the  

European  Economic  and  Social  Committee  and  the  Committee  of  the  Regions  SWD  (2017)  2  Final,  
COM(2017)  9  Final,  13;  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  
the  European  data  economy,  COM  /2017,  9  Final,  10.1.2017,  33,  making  reference  to  the  the  works  of  
Herbert  Zech,  that  have  claimed  that  the  right  forward  is  the  creation  of  a  property  right  to  non-­‐personal  
goods.  Cf.  H.  Zech,  Information  as  a  tradable  commodity,  in:  De  Franceschi  (Ed.),  European  Contract  Law  
and  the  Digital  Single  Market,  2016,  pp.  51-­‐79;      
50  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  

economy,  COM  /2017,  9  Final,  10.1.2017,  33.  

51  Ibid,  34.    
      15  

when the data has arrived in the database, the issue of authorship to the database
would, presumably, not be a problem anymore.

Indeed, if the main problem concerns the non-personal data flow between device and
database, a purely defensive right, if any, to the data is to be preferred.52 Any trading
would be conducted by giving license to the database.

Notwithstanding the above, it should be pointed out that an author of a database

including personal data still holds a copyright protection to the database, so possible
a defensive right also to personal data could be imagined. A defensive right in the
data, and a property right (copyright) to the database, would not directly intrude on
the privacy and personal data protection rights held by the subject of the personal
data.

From the general, brief, analysis above, even though possibly being a great source
or wealth for the future, it is rather clear that Data, be it Big or Open, personal or not,
is not directly covered by any intellectual property legal system, while still heavily
embedded in intellectual property rights protection and/or by neighbouring legal
protection. Indeed, the Digital economy, the technical aspects of IoT are, generally,
protected by patents and copyright protected software, the investment in the
collection of data may be protected by the sui generis data-base right and the data in
aggregate may be a trade secret, while the data subject still holds some rights to it
according to the General Data Protection Regulation (e.g. right to be forgotten and,
possibly, data portability).53 Indeed, the high amount of intellectual property rights in
the digital economy begs two general conclusions. Firstly, before creating new forms
of property rights in data, as such, the market and the industry must be analysed in
great detail. Do the industry benefit from property rights in data, or will it contractually
be able to sort out any issues of controversy. Secondly, the technical standards
seem to be inevitable in a digital economy with a great amount of intellectual property
and where, as a major aspect for success, interoperability between devices and
systems is of outmost importance.
                                                                                                               
52  W.  Kerber,  A  New  (Intellectual)  Property  Right  for  Non-­‐Personal  Data?  An  Economic  Analysis,  GRUR  Int  

2016,  p.  989,  appears,  also  to  the  Commission    to  be  favouring  such  an  approach.      
53  Cf.  Art  18-­‐20  GDPR.  The  rights  to  personal  data  should  however  be  weighted  against  the  other  rights  

acknowledge  by  the  Charter.    

      16  

4. Standards for the Internet of Things, Industrial Internet – the Issue of

Interoperability

4.1 Basic infrastructure interoperability standards and Upper Layer Standards

There is a great need for interoperability so to create the digital economy. Basic
interoperability, to materialize, needs some joint way of communication, or that one
technology is adopted as the technical standard either de facto or de jure by the
industry as a whole. Indeed, for the digital economy, IoT standards in general and
functional IP rules/guidelines under these standards in particular seem to be
desirable. When IoT will happen, devices will communicate with other devices, with
the telecom technology, and with the Cloud. Devices and device producers need
interoperability otherwise the system may not materialize. We are at the beginning of
the development of everything’s interoperability.

There is currently a global SSO race for IoT. Several different SSOs are fighting to
become the SSO part of the collaborations that enact the standards for the new IoT
era. Moreover, several pre-standard collaborations (consortia) are being formed
including several different combination of important players for the technologies that
might be included in the IoT standards. These consortia are like pacts conducting
lobbying and outright frontal attacks on other formations or pacts, all in the effort of
getting the ’right’ technologies inside the relevant standards.54

Of course, the incumbent SSOs are adapting to the new paradigm, e.g. ETSI, even
CEN and CENELEC claim relevance in the IoT paradigm.55 But there are special IoT
SSOs. For example, in 2015, the Commission and various IoT players launched a
large scale alliance called AIOTI (Alliance for Internet Of Things Innovation) with the
aim to assist the European Commission in the innovation and standardisation
policies.56

                                                                                                               
54  For  a  anaysis  of  the  political  /  techncial  interface  of  standard-­‐setting  cf.  Björn  Ludnqvist,  

”Standardization  under  EU  Competition  Rules  and  US  Antitrust  Laws  –  The  Rise  and  Limits  of  Self-­‐
Regulation,  Edrward  Elagr  2014,  16  et  seq.    
55  Cf.  CEN  and  CENELEC  (2016),  3  et  seq.  
56  See  conference  paper  by  Ciani,  (2016).    
      17  

Whether these collective efforts under SSOs will be relevant in a world where Google
(Brillo and Weave), Apple (HomeKit), Samsung (SmartThings), Amazon (Alexa) and
Microsoft (Windows 10 IoT editions) are all bringing out their own unilateral IoT
solutions is still to be seen. Perhaps, there will be no market for the SSOs and these
firms will instead create the de-facto/de jure IoT standards either just for their
respective ecosystem (including many vertical dependent firms) just as Google’s
Android became the de-facto open mobile OS?57

There are several layers of technical standards, and only a few standards are
actually of infrastructure interoperability type. Indeed, as Nicolo Zingales, stated, at a
minimum, one should distinguish standards for the “lower” and the “upper” layer,
pointing to a division between infrastructural interoperability and data interoperability.
While infrastructural interoperability enables devices to exchange data under
common network protocols, data interoperability concerns more directly users and
developers of IoT applications, allowing them to meaningfully connect mainly their
software interfaces of those applications. 58 Indeed, upper layer interoperability is
attained by reading and reproducing specific parts of computer programs, called
interfaces, which contain the information necessary to “run” programs in a compatible
format.59

Nicolo continues, it is from the perspective of the user/consumer of the computer

program, user interfaces are relevant to the extent that they enable him or her to
visualize and deploy a specific set of commands or modes of interaction with the
program, that can potentially be replicated into another (different) application.
Importantly, although this kind of interoperability can increase a program’s utility to
the user, it is not required for the purpose of its technical functioning. Most choices
for user interfaces are indeed dictated not so much by functional elements of the
program, as by the pursuit of the goals of user friendliness, aesthetical appeal and

                                                                                                               
57    Hughes,  ’A  world  with  more  IoT  standards  bodies  than  IoT  standard’,      

http://internetofthingsagenda.techtarget.com/blog/IoT-­‐Agenda/A-­‐world-­‐with-­‐more-­‐IoT-­‐standards-­‐
bodies-­‐than-­‐IoT-­‐standards.  (last  visited  23  December  2016),1  et  seq.  
58  Zingales,  Nicolo,  Of  Coffee  Pods,  Videogames,  and  Missed  Interoperability:  Reflections  for  EU  

Governance  of  the  Internet  of  Things  (December  1,  2015).  TILEC  Discussion  Paper  No.  2015-­‐026.  
Available  at  SSRN:  https://ssrn.com/abstract=2707570.  
59  Ibid.  
      18  

promotion of brand-specific features.60 Moreover, there does not seem to be one way
of solving the need for upper level interoperability since communication will be done
in several modes. Indeed, it seems that different collaboration of firms, or
ecosystems, may provide similar but still different designed data interoperability.

In addition, there also seem possible to make a distinction between horizontal and
vertical upper level standards. Indeed, as Kerber and Schweitzer explain,
”[p]articularly important is the distinction between horizontal and vertical [sematic]
interoperability. Horizontal interoperability denotes the interoperability of competing
products, services or platforms. One example is the interconnection between
communication networks. Vertical interoperability refers to the interoperability of a
product, service or platform with complementary products and services. The degree
to which complementary products (e.g., digital goods as music files or e-books) can
be shared across different platforms, and complementary products of one platform
can be accessed from rival platforms is said to characterize the horizontal openness
of a platform. The ability of independent firms to offer complementary products on a
platform stands for its vertical openness.”61

This is an interesting observation, and an issue would be whether, from a competition

policy perspective, horizontal openness in vertical interoperability should be
promoted, or whether system competition, between ecosystems, should be
promoted. Kerber and Schweitzer seem to urge caution in creating open (industry-
wide) standards in this regard, and that competition between vertical interoperability
systems (ecosystems) should be given room to foster the market. Only when we can
identify market failure should industry wide standards be promoted, and tolerated,
under competition law. Industry-wide standards may have several anticompetitive
effects, especially by lessening or totally eliminating competition between technical
systems (i.e. substitute technology competition), while only allow for competition
between complementary products in the same technology (rivalry based on the same

                                                                                                               
60  Ibid.    
61  Kerber,  Wolfgang  and  Schweitzer,  Heike,  Interoperability  in  the  Digital  Economy  (January  31,  2017).  

Forthcoming  in:  Journal  of  Intellectual  Property,  Information  Technology  and  Electronic  Commerce  Law  
(Jipitec);  MAGKS,  Joint  Discussion  Paper  Series  in  Economics,  No.  12-­‐2017,  4.  Available  at  
SSRN:  https://ssrn.com/abstract=2922515  or  http://dx.doi.org/10.2139/ssrn.2922515  
      19  

technology platform, through competition based on design and some form of

imitation).62

The two main market failures Kerber and Schweitzer identify are, firstly, dominant or
monopoly firm, where one firm won the rivalry for the demand and obtained unilateral
power to decide on the prevailing standard. The authors put in this group the
monopolist that obtained its market power through network effects and “tipping”. The
second example of market failure that requires an industry-wide standard is
according to the authors when the technology being standardized is of infrastructure
character an the economic advantages are so great that the industry can only have
one standard, i.e. the market tend to be a natural monopoly.63

Presumably, a third market failure situation would be markets where the patent or
intellectual property thicket has become so wide that the industry or relevant market
is not functioning or even emerging; and that access to SEPs need to be obtained
through a industry wide de jure standard with the complementary SSO IP
Guidelines.64

It is interesting to pursue the idea that industry-wide standards should be leniently

treated or even promoted in these three instances of market failure, while in other
scenarios consortia driven standard-setting as well as unilateral standard efforts
should benefit from heighten scrutiny under competition law since such collaboration
may lead to exclusionary collusive behaviour or anticompetitive exclusionary abusive
conduct. This issue will be more discussed infra.

Unfortunately, as Kerber and Schweitzer point out, the EU’s policy with regard to
collective standard-setting is not limited to privileging and supporting market-driven
cooperative standard-setting endeavours as a “bottom-up” approach. According to
the authors, the EU Commission, being concerned that, at least in the ICT sector,
standardization is increasingly taking place outside of Europe, potentially
undermining European competitiveness, it finds that it cannot be left to industry
                                                                                                               
62  ibid.,  4  et  seq.    
63  Ibid.,  7  et  seq.    
64  Björn  Ludnqvist,  ”Standardization  under  EU  Competition  Rules  and  US  Antitrust  Laws  –  The  Rise  and  

Limits  of  Self-­‐Regulation,  Edrward  Elagr  2014,  16  et  seq.  

      20  

stakeholders to decide in which areas to develop standards, and at what speed.

Rather, the Commission is, according to Kerber and Schweitzer, determined to
“define missing technological standards that are essential for supporting the
digitisation of our industrial and services sectors” and to actively mandate European
standardisation bodies for a speedy delivery of standards in order to “ensure that
ICT-related standards are set in a way that is more responsive to policy needs” and
sufficiently fast.65

In contrast, competition authorities should be more cautious vis-a-vis standard

consortia and other forms of pre-market collaborations now being set up for the
upcoming Internet of Things/Industrial Internet paradigm then the establishment of
dominance in the Data industry. Are we seeing signs of firms trying to agree on
technical solutions not so to facilitate Internet of Things, but rather to promote their
technical solution, while excluding others’ technical solutions? 66 Aspects of such
conduct might be procompetitive, while other aspects might be anticompetitive. The
competition authorities are thus forced to weight benefits of such collaborations vis-à-
vis their negative impact on competition. Indeed, the legislators’ as well as the
Commission’s effort should be to create a vibrant competitive Data industry
                                                                                                               
65  Kerber,  Wolfgang  and  Schweitzer,  Heike,  Interoperability  in  the  Digital  Economy  (January  31,  2017).  

Forthcoming  in:  Journal  of  Intellectual  Property,  Information  Technology  and  Electronic  Commerce  Law  
(Jipitec);  MAGKS,  Joint  Discussion  Paper  Series  in  Economics,  No.  12-­‐2017,  13  et  seq..  Available  at  
SSRN:  https://ssrn.com/abstract=2922515  or  http://dx.doi.org/10.2139/ssrn.2922515.  According  to  the  
recently  published  ICT  standardisation  priorities  (EU  Commission,  ICT  standardisation  priorities  for  the  
Digital  Single  Market,  Brussels,  19.4.2016,  COM(2016)  176  fin.,  EU  Commission,  Communication  “5G  for  
Europe:  An  Action  Plan”,  Brussels,  14.9.2016,  COM(2016)  588  fin.:  ”A  lack  of  coordination  between  
national  approaches  would  “create  a  significant  risk  of  fragmentation  and  implementation  of  standards  
and  would  delay  the  creation  of  a  critical  mass  for  5G-­‐based  innovation  in  the  Digital  Single  Market”  (p.  3).  
The  EU  Commission  finds  that  “standards  are  of  paramount  importance  to  ensure  the  competitiveness  and  
interoperability  of  global  communication  networks”  (p.  7)  and  plans  to  “foster  the  emergence  of  global  
industry  standards  under  EU  leadership  for  key  5G  technologies  (radio  access  network,  core  network)  and  
network  architectures”  (p.  7)),  open  European  standards  for  5G  communications,  for  the  IoT,  for  
cybersecurity,  big  data  and  cloud  computing  will  be  core.  In  various  areas,  the  new  digital  economy  
requires  an  “open  platform  approach  that  supports  multiple  application  domains  and  cuts  across  silos”.  
Open  standards  shall  support  the  entire  value  chain  and  integrate  multiple  technologies  (p.  7).  In  
particular,  the  Commission  is  interested  in  such  open  platforms  and  standards  in  the  area  of  eHealth,  
transport  systems,  including  automated  vehicles,  smart  energy  and  advanced  manufacturing  (p.  10  et.  
seq.).  At  the  same  time,  the  new  standardisation  processes  shall  take  into  account  the  blurring  of  the  
boundaries  between  traditional  sectors  and  industries,  products  and  services.  They  shall  consider  safety  
needs,  data  exchange  and  privacy  concerns  simultaneously  (p.3)  –  aspects  that,  today,  are  typically  dealt  
with  separately.  In  this  perspective,  the  Commission’s  pro-­‐collective  standard-­‐setting  approach  is  not  
limited  to  addressing  market  failures.  Rather,  what  resonates  in  these  communications  and  statements  is  
European  standard-­‐setting  is  a  pro-­‐active  trade  and  industrial  policy.”  See  also  EU  Commission,  A  Digital  
Single  Market  Strategy  for  Europe,  Brussels,  6.5.2015,  COM(2015)  192  fin.,  p.  15.  
66  Justus  Baron  and  Tim  Pohlmann,  ’Who  Cooperates  in  Standard  Consortia  –  Rivals  or  Complementors’  

(2013)  9  Journal  of  Competition  Law  &  Economics,  905.  

      21  

environment in an IoT setting, but the Commission, and other Competition Agencies,
needs to work ‘smart’ and focus on collaborative and unilateral conduct that are likely
to become anticompetitive when the markets of IoT can be identified.

4.2 Interoperable, standardized, secure and open-access platforms

Interestingly, creating interoperability (infrastructure or data interoperability)

standards so that devices may communicate perhaps is not the greatest of
concerns.67 On the contrary, in the IoT discussion the call for standardization seems
sometime, instead, to be a call not only for a joint standard for communication
between devices, but for access to data. Indeed, even the Commission recently
seems to suggest that access to data, or a wider re-usability of data, including
privately held data by other economic player could be a way forward so to create a
digital economy.

The Commission and the OECD in this debate make reference to a "data commons"
as a way to describe non-discriminatory access to certain data for at least a wider
group of players. According to these organisations, this should neither be confused
with an "open data" or "open access" approach (access for the public at large), nor
should it mean that access is given at no costs. The defining element of a
"commons" is that non-discriminatory access is to be given, i.e. any member of a
certain group, e.g. users of an industrial data platform, can use the data for purposes
defined by the party making the data accessible.68

Interestingly, several of the sector specific regulations that have been introduced in
the EU regarding data actually concern creating standards, not for interoperability,
but for access. Sector specific regulations seem to be the tool to be used to access
competitors data.

The PSI Directive already mentioned, stipulates route for accessing government
data. The main focus of the PSI Directive is very specific. It is to create a leveled
                                                                                                               
67  Only  perhaps  they  way  you  do  it.    
68  Cf.  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  

economy,  COM  /2017,  9  Final,  10.1.2017,  37.  See  also  OECD,  Maximising  the  Economic  and  Social  Value  of  
Data  (forthcoming).      
      22  

playing field when making available PSI as input to a commercial activity, i.e. when
the PSI is used as components to new products and services. This should then
release the full economic potential of a new emerging area of the ICT sector.69 If
PSBs offer information products or services exclusively, chances are that they will
not be able to provide these services as innovatively and efficiently as a structure
governed by competition would be able to.70 This could have a negative effect on
competition in the European market. Therefore, the PSI Directive aims to overcome
these barriers, which limit the re-use of PSI in EU Member States. The PSI Directive
thereby stipulates public sector data collectors should grant access to data. The
public sector data collector is even obliged to grant access if it re-use data collected
commercial by selling access to the data base to data brokers or re-users. The PSI
Directive tries to negate barriers which could include attempts by Public Sector
Bodies (PSBs) to charge supra competitive prices, unfair competition between the
public and the private sector, practical issues hindering re-use (like the lack of
information on available PSI), and the attitude of PSBs failing to realize the economic
potential of PSI.71 The PSI Directive is triggered by three sets of questions:

1. Are the data created (supplied) outside or inside the public task of the Public
Sector Body? If so, what is the initial purpose for producing the data? Is it to fulfill a
public task? If yes, the second question is:
2. Are the documents being re-used by the PSB or some other body on its behalf? In
other words, will the data be used for another purpose than the initial purpose?
Moreover, will this re-use constitute a commercial activity, e.g. giving access to the
dataset to paying subscribers?
3. If so, then a number of requirements will apply, including that third parties have the
right to access the dataset on something similar to FRAND terms, so to enable the
third parties to commercially utilize the Open Data in competition with the PSB and
other firms having access to the dataset.

                                                                                                               
69  EU  Commission,  (1998),  5.  
70    Lundqvist  Vries,  Linklater  and  Rajala  Malmgren,  ‘Business  Activity  and  Exclusive  Right  in  the  Swedish  

PSI  Act’,  Report,  Swedish  Competition  Authority,  (September  2011),  17.  

71  Janssen  and  Dumortier,  “Towards  a  European  Framework  for  the  Re-­‐use  of  Public  Sector  Information:  a  

Long  and  Winding  Road”,  2  International  Journal  of  Law  and  Information  Technology  (2011),  195.  
      23  

Interestingly, the PSI directive seems to include a non-discriminatory exclusion rules,

similar to the ideas put forward in the French competition law cases GDF72 and
EDF73 discussed below.

The very interesting eCall Regulation 74 should also be mentioned. According to

Recital 16, “[i]n order to ensure open choice for customers and fair competition, as
well as encourage innovation and boost the competitiveness of the Union's
information technology industry on the global market, the eCall in-vehicle systems
should be based on an interoperable, standardised, secure and open-access
platform for possible future in-vehicle applications or services. As this requires
technical and legal back-up, the Commission should assess without delay, on the
basis of consultations with all stakeholders involved, including vehicle manufacturers
and independent operators, all options for promoting and ensuring such an open-
access platform and, if appropriate, put forward a legislative initiative to that effect.
Furthermore, the 112-based eCall in-vehicle system should be accessible for a
reasonable fee not exceeding a nominal amount and without discrimination to all
independent operators for repair and maintenance purposes in accordance with
[…][author’s highlight].75

The eCall directive seems to lay the groundwork for a future where competitors are
able to access data originating from cars used by individuals. The groundwork being
that the devise should be (connected to) a standardized, secure and open-access
platform. Interestingly, the idea seems to be that by creating a standard, competitors
will be enabled not (only) to produce eCall machines under FRAND licenses, but to
actually access eCall machines in the cars with their own applications so to pick up
Data. The automobile manufacturers should, thus, not have exclusive right to the
personal data created in the car (the device), while possibly, be open up the platform
in the car to, for example, leasing firms, insurance companies and independent

                                                                                                               
72  French  Competition  Authority,  Decision  14-­‐MC-­‐02  of  09.09.2014.  The  case  is  discussed  in  the  joint  

report  by  the  German  and  French  Competition  Authorities,  (2016),  20.    
73  French  Competition  Authority,  Decision  n°13-­‐D-­‐20  of  17.12.2013,  confirmed  on  that  points  by  the  court  

of  appeal  on  21.05.2015.  

74  Regulation  (EU)  2015/758  of  the  European  Parliament  and  of  the  Council  of  29  April  2015  concerning  

type-­‐approval  requirements  for  the  deployment  of  the  eCall  in-­‐vehicle  system  based  on  the  112  service  
and  amending  Directive  2007/46/EC.  
75  Ibid.  
      24  

service providers should be able to access the device for collecting data.76 Indeed,
Commission in its latest report states that preparing for a legislative initiative on "an
interoperable, standardised, secure and open-access platform", the issue of access
to in-vehicle data has been discussed in depth since 2014, in the framework of a C-
ITS stakeholder platform. Five guiding principles for access to in-vehicle data and
resources have been agreed upon, including fair and undistorted competition and the
importance of "standardised access" to in-vehicle data as an enabler of the "common
use" of such data in the context of the "data economy". Work is on-going in terms of
evaluating the technical and legal feasibility of improving access to in-vehicle data.77

Finally, the recently updated Directive on Payment Services stipulates a right for third
parties under certain circumstances to access the banking data of consumers.
Consumer should be able to agree that third parties provide services, while
accessing consumer bank accounts and internet bank sites. DPS II may, to promote
competition, require banks to provide standardized API access to third parties under
the auspices of the European Banking Authority (EBA). 78 This may enable third
parties to tailor their banking service towards customers, while using data collected
by a competitor.

These three legal acts are examples of rules that are, or stipulating that there soon
will be rules, requiring competitors to give access to personal and non-personal data
or devices/platforms so to enable data mining. It is possibly an indication of a
interesting underlying current that the legislator try to boost competition by granting
access to competitors data, while circumventing general competition law. The idea is
to boost competition, without making use of any test of antitrust harm, by opening up

                                                                                                               
76  An  example  of  this  development  could  be  the  AUDI  AG,  BMW  Group,  Daimler  AG,  Ericsson,  Huawei,  

Intel,  Nokia  and  Qualcomm  Incorporated,  announce  27  September  2016  the  formation  of  the  “5G  
Automotive  Association”.  The  association  will  develop,  test  and  promote  communications  solutions,  
support  standardization  and  accelerate  commercial  availability  and  global  market  penetration.  The  goal  is  
to  address  society’s  connected  mobility  and  road  safety  needs  with  applications  such  as  connected  
automated  driving,  ubiquitous  access  to  services  and  integration  into  smart  cities  and  intelligent  
transportation.  https://www.ericsson.com/news/160927-­‐telecommunications-­‐and-­‐automotive-­‐
players_244039854_c.  
77  Commission  Staff  Working  Paper  on  the  free  flow  of  data  and  emerging  issues  of  the  European  data  

economy,  COM  /2017)  9  Final,  10.1.2017.,  25.  

78  See  Directive  (EU)  2015/2366  of  the  European  Parliament  and  of  the  Council  of  25  November  2015  on  

payment  services  in  the  internal  market,  amending  Directives  2002/65/EC,  2009/110/EC  and  
2013/36/EU  and  Regulation  (EU)  No  1093/2010,  and  repealing  Directive  2007/64/EC  (Text  with  EEA  
relevance).  Cf.  Commission  (2015).    
      25  

the device for collecting data to all-and-everyone. Whether such a policy is pro-
competitive may be disputed. Not only new entrants will be able to obtain data, also
incumbent e-platform firms will try to access these devices or, in reference to
government data, the PSI. This may, for example, act as a deterrent for the brick-
and-mortar industry firms to become full-fledged competitors in the Data industry.
Indeed, the incentive of becoming members of the data industry may be low if a
brinck-and-morter firm knows that it is obliged to share the input Data, i.e. its raw
material. Moreover, it also begs the question, to be put to the legislator, why similar
access rules do not exist in other industries. Why not for e-platform providers as such
when large industries such as the car industry and the bank sector should be
required to give access?

5. The Application of Competition Law

5.1. Standard-setting – so to create interoperability

According to the Commission, where participation in standard-setting is unrestricted

and the procedure for adopting the standard in question is transparent,
standardization agreements, which contain no obligation to comply with the standard
and provide access to the standard on fair, reasonable and non-discriminatory terms,
will not normally restrict competition within the meaning of Article 101(1) TFEU.79 The
paragraph indicates that there is a safe harbour for Standard Setting Organizations
(SSOs) (and its members), which is described in the horizontal guidelines, published
by the EU Commission (Horizontal Guidelines). It caters to all SSOs, public, semi-
public and private, and, also, irrespective the SSO and its members have market
power, or not. Indeed, the safae harbour is alos applicable to structures with market
power. Under US antitrust law, the seldom discussed, but important, National
Cooperative Research and Production Act of 1993 (NCRPA) was amended in 2004
by the Standards Development Organization Advancement Act, in order to extend
the protection from the antitrust laws to specified activities of SSOs, which had
previously been granted to R&D collaborations. If an SSO discloses the scope and

                                                                                                               
79  Guidelines  on  the  application  of  Art  101  of  the  Functioning  of  the  European  Union  to  horizontal  co-­‐

operation  agreements,  OJ  C  11,  14/01/2011,  1  (Horizontal  Guidelines).  

      26  

nature of its activities to the Department of Justice, it may take advantage of the
NCRPA.

Under the legislation described above, technology standardization agreements and

SSOs are, irrespective of market power, de facto exempted or per se legal, i.e.
immune under EU competition law and US antitrust rules as long as the rules of the
safe harbours, i.e., procedure, good governance and transparency etc. stipulated in
either the Horizontal Guidelines or the NCRPA are adhered to. Indeed, even if the
consortia is not open to all comers, the joint standard-setting initiative will very likely
either fall outside the antitrust regulation or be exempted. 80 That joint standard-
setting is so general exempted without a more sophisticated analysis of the
collaboration and, foremost, without an analysis of the technology and market being
standardized is unlucky. There are certainly joint standard-setting efforts that should
be exempted from the antitrust rules, others that should be exempted, but there is
also collaborations that might benefit from a more intense antitrust scrutiny. Joint
standard-setting can be conducted also with the effort to exclude innovations and
competitors, increase prices and look-in firms in certain technologies.

Notwithstanding the above, it seems that there are clear arguments for why neither
Schumpeter’s theories of dynamic competition, nor Hayek’s concept of competition
as a discovery procedure, may apply for certain standard-setting processes or for
some forms of markets.81

It seems that their ideas of rivalry and competition are based on two assumptions.
First, as Drexl states ‘the functioning of Schumpeterean competition relies on
particular assumptions that correspond to the economic assumptions of traditional
patent law. In particular, it is assumed that each product competing in a relevant
market is protected by a single, discrete patent. In this ‘ideal world of Schumpeterean
competition’, patent-holding manufacturers only compete in product markets and use
their patents as instruments for protecting investment in better products against

                                                                                                               
80  Extensivelly  developed  in  Björn  Lundqvist,  Standardization  under  EU  Competition  Rules  and  US  Antitrust  

Laws    (Edward  Elgar,  2014).  

81  Cf.  generally  Friedrich  A.  Hayek,  New  Studies  in  Philosophy,  Politics,  Economics,  and  the  History  of  Ideas  

(Routledge  and  Kegan  Paul,  1978),  185  et  seq.  Joseph  Schumpeter,  Capitalism,  Socialism  and  Democracy  
(George  Allen  &  Unwin  1976  (first  published  in  1943)),  82  et  seq.  
      27  

imitation for free.’82 Second, both Schumpeter and Hayek assume that the market in
question is a functioning market where different products/technologies may compete
for the benefit of customers. In todays new economy neither of these assumptions
will always hold true. Firstly, there is patent mosaic in certain markets where the
relevant market is covered by several patents held by several firms, and, second,
there are markets that may not bear competition between different solutions for a
demand. Infrastructure technologies and markets plagued by network effects will
likely develop an omnipresent standard, while they may furthermore be considered
natural monopolies or other wise ‘failed’ markets where competition by substitute
technologies need to be sacrificed for benefits of competition by design or even
imitation.

As stated above, standards today reflect fundamentally different market solutions.

Some standards represent infrastructure, others interoperability solutions, which will
de facto create not only new downstream markets but possibly are the new
‘structures’ for these markets, while other standards reflect the joint effort (or
ecosystem) of firms to sell products and services in existing relevant social site
markets. These ecosystems may have a firm with great power within the ecosystem,
the system leader, while still allow vertical access by allowing independent firm to
produce applications for the ecosystem. Indeed, the strong role of platform markets
(search engine market, social media market etc.) with their strong positive network
effects and tipping may create quasi-monopolists.83

However, implementing an infrastructure standard implies selecting one solution for

an industry or market. Such infrastructure standards are not developed through a
‘trial and error’ method in which the consumers are exposed to different solutions.
Infrastructure standards are rather developed through a standard-setting process
where the relevant parties negotiate and agree on the infrastructure while conducting
R&D during the period the standard is developed. These standards should, therefore,

                                                                                                               
82  Josef  Drexl,  ’Intellectual  property  in  competition:  How  to  promote  dynamic  competition  as  a  goal’  in  

Josef  Drexl  et  al  (eds),  More  Common  Ground  For  International  Competition  Law?  (Edward  Elgar,  2011),  
211  et  seq.  
83  Haucap/Heimeshoff,  Google,  Facebook,  Amazon,  eBay:  Is  the  Internet  driving  competition  or  market  

monopolization?,  Int  Econ  Econ  Policy  2014,  49,  50  et  seqq.  Evans,  suggests  that  tipping  towards  
monopolies  is  usually  prevented  by  the  complexity  of  multi-­‐platform  markets:  The  Antitrust  Economics  of  
Multi-­‐Sided  Platform  Markets,  Yale  Journal  on  Regulation,  Vol.  20  (2003),  325,  350.      
      28  

not be considered the same kind of goods and services which Hayek, and for that
matter, Schumpeter were discussing. In many ways, these authors discuss, even
though in the abstract and theoretically, goods and services over which consumers
and society de facto have a choice; where the market has not failed.84 In other words,
where the consumer may select a good, consume it or try it out, and purchase
another good if the first good did not meet their satisfaction. Thus, where there is no
market failure and the fundamental principles of Hayek, Schumpeter and Arrow work
(i.e. that there is a choice), the innovation or competition process through ‘trial and
error’ may be applied. However, in today’s ‘high-tech’ industries, certain
technologies, which are currently being standardized, are more akin to infrastructure
than to goods. The reason for this is that these technologies are often plagued by
network effects and customers/consumers are locked in when the technology is
selected. As Drexl states, ‘[m]arkets characterized by network effects will
unavoidably migrate towards technological standards. The market position gained by
the person controlling the standard based on IPRs is therefore to be considered a
natural monopoly, which is economically superior to forced competition.’ 85 Thus,
society needs to agree, de jure, on, or accept, de facto, a standard, and it is unlikely
that the parties to these standards, including suppliers, content providers and
consumers, will be able to back-track when the technology or standard is selected.
They are, even though sometimes short-lived, still infrastructure or even natural
monopolies. For these standards, competition as a process, unfortunately, does not
suffice as a theoretical groundwork to select the best technology for society. First, a
‘trial and error’ method could be a very costly method for selecting infrastructure.
Secondly, if these standards are not agreed upon, de facto standards due to network
effects will soon emerge. However, de facto standards created through network
effects and ‘tipping’ could very well be inferior to a de jure (i.a. agreed upon)
standard. Due to the network effects and tipping, a ‘trial and error’ method would
imply that a less good solution would possibly become the de facto standard.

                                                                                                               
84  See  Drexl  for  a  discussion  regarding  external  market  failure  and  competition  law.  Josef  Drexl,  'Abuse  of  

Dominance  in  Licensing  and  Refusal  to  License:  A  More  Economic  Approach  to  Competition  by  Imitation  
and  to  Competition  by  Substitution',  in  Claus-­‐Dieter  Ehlermann  and  Isabela  Atanasiu  (eds),  European  
Competition  Law  Annual  2005:  The  Interaction  Between  Competition  Law  and  Intellectual  Property  Law  
(Hart  Publishing,  2007).  652  et  seq.    
85  Josef  Drexl,  ’Intellectual  property  in  competition:  How  to  promote  dynamic  competition  as  a  goal’  in  

Josef  Drexl  et  al  (eds),  More  Common  Ground  For  International  Competition  Law?  (Edward  Elgar,  2011),  
212  et  seq.  
      29  

Instead, the parties must be allowed, under dynamic competition, to not only choose
between available technologies, but also discuss the pros and cons of different
technical solutions, while individually continuing to conduct R&D efforts. In these
circumstances, the standard-setting process must be understood as a form of
collective innovation and also collective competition, where the end result is a
standard that meets the demands of society as a whole.86 Under such standards
there will, however, only be room for competition by imitation, while competition by
substitution will not suffice due the inherent market failure of creating one
omnipresent standard.87

On the other hand, where there is room in the market for consumers to choose
between different solutions to a problem, firms must have the ability to compete for
the demand. Therefore, standards or technology which represent data interoperability
within an ecosystem, which in turn run on a given syntactic/technical interoperability
technology, should be developed through a ‘trial and error’ method, i.e. be exposed
to consumers. In these settings, different forms of consortia should compete with
different data interoperability/solutions and standards. Here horizontal interoperability
for a vertical ecosystem may be a competitive advantage.

Thus, a standard such as the one reflecting 3G telecommunications technology

should be dealt with differently under competition law, than ‘data interoperability
standards’. Under competition law, we must tolerate more collaboration between
competitors when creating de jure or de facto standards reflecting infrastructure or
syntactic/technical interoperability, while opposing firms that try to inhibit or hold up
such standards from being implemented properly. Thus, competition by imitation
under such standards should be promoted, while competition by substitution will not
be present.

On the other hand, a product standard on a market without network effects that
reflects one technology out of several technologies in a market can create anti-
                                                                                                               
86  Josef  Drexl,  'Anti-­‐competitive  Stumbling  Stones  on  the  Way  to  a  Cleaner  World:  Protecting  Competition  

in  Innovation  without  a  Market',  (2012)  8  Journal  of  Competition  Law  &  Economics,  507,  534.    
87  Josef  Drexl,  'Abuse  of  Dominance  in  Licensing  and  Refusal  to  License:  A  More  Economic  Approach  to  

Competition  by  Imitation  and  to  Competition  by  Substitution',  in  Claus-­‐Dieter  Ehlermann  and  Isabela  
Atanasiu  (eds),  European  Competition  Law  Annual  2005:  The  Interaction  Between  Competition  Law  and  
Intellectual  Property  Law  (Hart  Publishing,  2007),  651  et  seq.    
      30  

competitive effects if that standard de jure or de facto excludes competition, inter alia
by locking vertically integrated participants into that standard, without the possibility
to choose between different ecosystems. Where there are competing technologies
on a relevant technology market, or where there are several future possible
competing products, the collusive character of a consortium deciding on one
standard should be the less tolerated if that were to imply that competition by
substitution had been eliminated or lessen. However, firms trying to ‘hold up’
technical solutions should be treated with more tolerance. Such ‘hold-up’ strategies
may reflect competition on the merits, i.e. normal conduct in the dynamic competitive
process. Thus, for these product standards, collaborations that decide what
technology will be implemented in markets and society, without de facto protecting
substitute competition in the downstream markets where different firms may present
their different technology alternatives to consumers, should be scrutinized more
carefully.88 If technologies, which represent different competing solutions/supplies for
a given demand, may coexist in a relevant market, the standardization agreement
would benefit from a heightened antitrust scrutiny. Such a standardization agreement
should only stipulate that a technology should be the omnipresent product standard if
considerable efficiencies can be attributed to such a decision.

That said, an additional problem is the patent thicket and the notion of anticommons.
This may also create market failures. Certain product standards, especially in the IT
sector, but also in parts of the telecom sector and the consumer electronics sector,
are plagued with patent thicket problems. These technologies are still not to be
considered infrastructure, or even technologies for interoperability, but they cannot
be effectively utilized without the ability to access standard essential patents. Thus,
certain markets and products will not emerge without collective action under one
standard or patent pool to counteract the patent thicket. Standardization plays here a
significant role in codifying fragmented technologies into one standard.89 For these
markets, IP arrangements under standard agreements and patent pools are

                                                                                                               
88  For  the  notion  of  contestable  markets,  see  Carl  Shapiro,  'Competition  and  Innnovation:  Did  Arrow  Hit  

the  Bull's  Eye?',  in  Josh  Lerner  &  Scott  Stern  (eds),  The  Rate  &  Direction  of  Economic  Activity  Revisted  
(University  of  Chicago  Press,  2012).    
89  Liguo  Zhang,  Standardization  and  Patent  Licensing  in  the  European  Union  (Oy  Nord  Print  Ab,  2012),  104  

et  seq.  
      31  

necessary in order to create effective markets and competition in downstream

markets.

The opposite holds true when discussing the standards and the markets connected
thereto, which are not plagued by neither network effects nor patent thickets, i.e.
product, and not performance, standards for, for example, cement and tyres. Such
product standard collaborations, such as joint R&D in the pre-standardization phase,
which create a common standard and IP arrangements under standards and patent
pools, should be scrutinized more carefully to establish whether they exclude
competition. If there is no network effect on the relevant market, an all-industry
inclusive standard should benefit from a higher antitrust scrutiny.

In other words, one could divide technologies and markets according to the anti-
competitive effects of standards and their connected collaborations into four
categories.

A. B. C. D.
Basic/Infrastructure Network/Interoperability Patent thicket plagued Product

Collaborations and unilateral conduct under these four categories of technology

deserve different antitrust treatments. An A standard is often derived from a
completely open inclusive standard-setting process, where the organizations inherent
collusive character must be tolerated, while open access to membership and results
are often granted and protected under competition law, since such standards create
markets and competition. B network/interoperability standards should receive similar
treatment but there should be a requirement that the network and tipping effects of
the technology and standard are demand-driven rather that created by the suppliers
of the standard (this issue is discussed infra in connection with the discussion of DVD
technology and the HD-DVD v. Blu-Ray war).

On the other hand, under a D standard, an all-industry inclusive standard-setting

effort would diminish competition. Design or product standards (D) are those
standards that mirror a service or product, e.g. cement or a specific tyre, or even a
certain website or search engine, which in fact is only a product among others on an
      32  

existing or soon to emerge relevant market. Standards that identify one product as
‘the standard’ on such markets may cause exclusionary anti-competitive harm. The
object behind such a standard can be collusion, to limit the number of products so to
enable a price cartel. Exclusion of competing products or technologies might occur,
or even be the aim of such a standard.90

For C standards, a decision on a standard may still be efficient. If there is a patent

thicket that prevents relevant downstream product markets from materializing, or
from working efficiently, a standard with connected IP arrangements or patent pools
may very well be the tool that enables firms to penetrate the patent thicket. Thus,
these forms of product or design standard could, under certain circumstances, be
exempted from competition law.

Several of the cases in which the courts have been involved revolve around product
or design standards (D) and attempts by members of the SSO to create a monopoly
technology. The US cases like Allied Tube and Radiant Burners, on the one side,
and IAZ, EMC and also Fr.bo, on the EU side, represent this line of case law quite
well regarding collective exclusion or boycott. 91 The test seems to include the
following steps: (i) the market will sustain several standards (competition by
substitution); (ii) deciding on one or more standard with its related terms of use de
facto excludes a competitor or competing standard/technology from the market if, for
example, the standard is elevated into a regulation; (iii) a plaintiff needs to identify a
potential anti-competitive effect on competition and/or parallel trade either by
showing misuse of the standard-setting process, or, in rare circumstances, an anti-
competitive intent in the setting up of the standard on behalf of at least some
competitors in the SSO; and, finally, (iv) when the plaintiff has shown the above (i.e.
exclusionary effect or anti-competitive intent), the defendants may exculpate
themselves, according to Hovenkamp, by demonstrating the superiority of the
                                                                                                               
90  Price  cartels  that  have  been  analysed  by  the  EU  Commission  and  Justice  Department  may  have  a  

standard-­‐setting  side.  See  e.g.  Joined  cases  96  to  102,  104,  105,  108  and  110/82  IAZ  and  others  v.  
Commission  [1983]  ECR  3369  and  Commission  Decision  in  Case  IV/35.691,  Pre-­‐insulated  Pipes,  OJ  L  24,  
30.1.1999,  p.  1.    
91  Radiant  Burners,  Inc.  v.  Peoples  Gas  Light  &  Coke  Co.,  364  U.S.  656,  659  et  seq.  (1961);  Allied  Tube  and  

Conduit  Corp.  v.  Indian  Head,  Inc.  486  U.S.  492,  500,  509  (1988).;  Case  C-­‐171/11  Fr.bo  SpA  v.  DVGW,  not  
reported  yet.;  Joined  cases  96  to  102,  104,  105,  108  and  110/82  IAZ  and  others  v.  Commission  [1983]  ECR  
3369;  Case  COMP/F-­‐2/38.401  EN  197-­‐1  Standard-­‐  EMC/European  Cement;  and  Case  T-­‐432/05  EMC  
Development  AB  [2010]  ECR  II-­‐01629.  
      33  

standardized technology. As Hovenkamp develops the argument, the courts should

be very reluctant to become embroiled in evaluating complex technical concepts, but
there are times when the courts are forced to do so.92

Product or design standards should often be viewed as performance standards. For

these forms of design or product standards, the way the standard is described can
furthermore create anti-competitive effects: (i) ‘overstandardization’, i.e. the breadth
and depth of the technology standard excludes new and competing technologies;93
(ii) ‘understandardization’, i.e. the technology is still regulated under the standard but
the breadth or depth of the standard is too narrow and grants an exclusive field of
use to certain intellectual property holders within the technology.94

The B standards are different, and here the courts need to hesitate and conduct an
analysis of the relevant markets and innovation competition. Interface or
interoperability standards are very close to infrastructural standards. They represent
markets where network effects can be very strong and may only hold one technology
or product. If there is a true interoperability standard, such standard-setting
processes ought possibly to be treated in the same manner as an infrastructure
standard. However, it is a very difficult analysis to conduct for the courts. Several of
the technologies that we today see as just one competing technology on a larger
relevant technical market, were considered to be both interopreability technology and
the only solution for a demand, when the pioneer technology was launched.

The web browser market was identified by several commentators as only holding one
software product, without any room for competition by substitute. Thus, only one
‘product’ could survive the competitive process, and it also represented the interface

                                                                                                               
92  Hovenkamp  et  al.,  IP  and  Antitrust:  An  Analysis  of  Antitrust  Principles  Applied  to  Intellectual  Property  Law  

(Aspen  Publishers,  2010),  §  35.4,  35-­‐35.  

93  Cf.  Dolmans  regarding  standardization  depth,  i.e.,  what  freedom  is  left  for  firms  to  develop  future  

improvements  under  a  technology  standard.  Maurits  Dolmans,  'Standards  For  Standards',  (2002)  26  
Fordham  International  Law  Journal,  163,  172  et  seq.  See  also  Marcus  Glader,  'Open  Standards:  Public  
Policy  Aspects  and  Competition  Law  Requirements',  (2010)  6  European  Competition  Journal,  611,  615  et  
seq.  
94  For  a  similar  but  different  list,  cf.  the  headings  of  Maurits  Dolmans’,  'Standards  For  Standards',  (2002)  

26  Fordham  International  Law  Journal,  163.    

      34  

between the operating system (even the PCs as such) and the internet. 95 In
hindsight, we now know that that was not true. There exist several web browsers
today used by consumers. However, somewhere in the development of this market, it
went from being a unilateral to multilateral technology solution market.

The CD technology when launched may perhaps be considered as interoperability

technologies, where the relevant market may be interpreted as being plagued by
network effects. Several standards on the internet relating to the transfer of meta-
data are about interoperability. But, what about the DVD and the Blu-Ray
technologies? Here the alleged network effects seem actually to have been produced
by the content and the technology providers. It seems that the different DVD and Blu-
Ray technologies represented different vertical collaborations (or ecosystems),
between different providers of content or technology. Legally, these partner
collaborations (consortia) were made up of exclusive cross-licence agreements. The
exclusivity of these agreements could have constituted a competition law issue, while
the technologies seem both to be able to be present on the relevant market should
these exclusive agreements be void.

Indeed, the network effect seems to have been an effect of exclusive agreements
between content providers and technology providers in the DVD and Blu-ray wars. If
these exclusive arrangements had been eliminated these two technologies could
have competed in parallel.

Possibly, the DVD and Blu-Ray collaborations were collusions to create dominant D
standards and should benefit from a heightened antitrust scrutiny. These forms of
technologies for holding content (music and film) are no different from sticks, MP3,
the technology supporting Netflix, or for that matter the technology supporting the
website, Pirate Bay.com! Perhaps, if the competition authorities had tried to uphold
competition between these two different technologies and joint ventures on the
relevant technology and product markets, so that content providers could have
launched on both systems, problems, such as are represented by Pirate Bay, would

                                                                                                               
95  The  Microsoft  case  displays  the  antitrust  application  to  a  de  facto  interoperable  standard.  United  States  

v.  Microsoft  Corp.,  253  F.3d  34  (D.C.  Cir.  2001)  (en  banc).    
   
      35  

not have materialized. The price of the technologies in question would in any case
possibly have been a lot less than they became.

Indeed, the B standards seem, generally, to be derived from standard consortia/joint

R&D ventures between vertical firm with complementary services. When the
technology is acquired, in an attempt to get other suppliers (e.g. content providers)
and consumers to accept these standards, they are presented with an SDO. Often
they represent an inventive leap but they are not basic infrastructural standards.
Thus, these standards may easily fool us into believing that they represent the only
solution, although these markets would often benefit from several standards
competing.

C standards are product standards but where the standard agreement is entered into
to create a relevant market and a product out of a patent thicket. In other words, a
joint effort where the standard is drafted so as to enable trade and competition by
creating a collaboration between all essential patentees. These should even be
allowed to become industry-wide collaborations, as long as the SSO and the
members are able to prove that there is a patent thicket problem. Examples of
standards for this are the Moving Picture Experts Group (MPEG) digital audio and
video compression specification standards including MPEG-2, MPEG-4, Thompson’s
MPEG-1, and Leyer-III (MP3). It should be noted that in the DVD Business Review
Letters, the parties claimed that there was a patent thicket in the technology, while
this seems less certain in hindsight. These C standards are in fact often horizontal
agreements for the creation of patent pools.

It is therefore, as Kerber and Schweitzer developed, very unfortunate that the EU

Commission in several guiding documents promote joint standard setting, and that
joint standard-setting also in private consortia format are de facto exempted from
Article 101 TFEU under the Horizontal Guidelines. Indeed, a more sophisticated
analysis of the consortia in data portability should be conducted before the paradigm
shift of Internet of Things, so that group exclusion of new innovators or maverick
firms, and/or quasi-monopolistic power, are not being implemented. Moreover, in the
analysis of whether some forms of collaboration must be found to be anticompetitive
the issue of decomilation according to the Software Directive, Trade Secret Directive
      36  

and the Unified Patent Court Agreement should be remembered. Several patent laws
also include an experimental exemption. There are narrow possibilities to be allowed
to conduct research and of decompilation of copyright or patent protected technology
so to enable interoperability.96 Indeed, these rights should perhaps be taken into
consideration when analysing whether a collaboration should be allowed or not, since
they represent an alternative route to collaborations for accessing technologies.
Indeed, these rights may create a competitor.

5.2 Dominance and Abuse

The strong role of platform markets (search engine market, social media market etc.)
with their strong positive network effects and tipping may create quasi-monopolists,
which control and holds market power in its ecosystem.97

What is important in the standard-setting race for the ioT is that firms with market
power (system leaders) honour the integrity of standard-setting procedure. In
markets where tipping and network effects are present and may be created and
promoted by system leaders by exclusively vertically integrate certain critical
services, system leaders should not be able to “game the system” so to be able to
hold there infrastructure interoperability technology outside the SSOs and the
FRAND commitments under Huawei, while being able to obtain access to competing
firms interoperability technologies under de jure standards. Presumably, hold-up, but
also other forms of business strategies that imply gaming the system in reference to
basic infrastructure standards may be considered as an abuse.

Moreover, within ecosystems, system leaders are quasi-monopolists, and if there is

low competition between ecosystems, the market power of system leaders within
respective ecosystem may increase. Possibly, dominance should be identified in
these circumstances through using other forms of market power analysis than the
relevant market test, such as value chain economic tests.
                                                                                                               
96  Directive  2009/24/EC  on  the  legal  protection  of  computer  programs,  Trade  Secret  Directive  2016/943  

of  8  June  2016,  OJ  2016  L  157/1.      

97  Haucap/Heimeshoff,  Google,  Facebook,  Amazon,  eBay:  Is  the  Internet  driving  competition  or  market  

monopolization?,  Int  Econ  Econ  Policy  2014,  49,  50  et  seqq.  Evans,  suggests  that  tipping  towards  
monopolies  is  usually  prevented  by  the  complexity  of  multi-­‐platform  markets:  The  Antitrust  Economics  of  
Multi-­‐Sided  Platform  Markets,  Yale  Journal  on  Regulation,  Vol.  20  (2003),  325,  350.      
      37  

Indeed, in the area of platform markets, with thereto connected vertically integrated
firms, we may see new forms and tests of dominance and abuses, which the
competition authorities need to adopt under the new paradigm of the Internet of
Things.

5.3 In reference to access to data

At least in reference to public data (so-called open data), competition law has been a
great source of inspiration as a way to regulate the interaction been authorities and
re-users. Cases like Magill 98 , IMS Health 99 and Microsoft 100 resemble and,
presumably, influenced the way the public sector information (PSI) directive101 has
been drafted. Indeed, the interface between the PSI legislation and general
competition law, especially the abuse of dominance doctrine in reference to (i) refusal
to supply, (ii) exclusionary abuses and even (iii) discriminatory exclusion have been
scrutinized by some national courts and competition authorities in conjunction with
claims fo breach of the PSI rules (the PSI Directive is discussed infra).102

However, the use of general competition law doctrines such as refusal to supply (or,
the exceptional circumstance doctrine) to gain access to datasets may be somewhat
problematic.103 Is the data holder dominant on a relevant market? In reference to e-
platforms, Google and Facebook have been accused of holding market power due
                                                                                                               
98  ECJ,  RTE,  ITP  &  BBC  v.  Commission,  C-­‐241/91  and  C-­‐242/91,  ECLI:EU:C:1995:98.    
99  ECJ,  IMS  Health  v.  NDC  Health,  C-­‐418/01,  ECLI:EU:C:2004:257  .  
100  GC,  Microsoft  v  Commission,  T-­‐201/04,  ECLI:EU:T:2007:289  
101  The  Directive  on  the  re-­‐use  of  public  sector  information  (Directive  2003/98/EC,  known  as  the  'PSI  

Directive')  entered  into  force  on  31  December  2003.  It  was  revised  by  Directive  2013/37/EU  which  
entered  into  force  on  17  July  2013.  
102  Lundqvist,  (2013),  80  et  seq;  Lundqvist,  (2011),  11.    
103  In  reference  to  Public  Sector  Bodies,  the  issue  has  been  whether  they  can  be  regarded  as  undertakings.  

Firstly,  the  data  holder’s  activities  with  the  data  need  needs  to  be  analyzed  in  order  to  establish  whether  
the  holder  in  an  undertaking  in  reference  to  Article  102  TFEU.  Is  the  activity  under  scrutiny  an  economic  
activity,  i.e.  a  commercial  activity,  conducted  on  a  market?  This  may  only  be  established  if  the  end  market,  
where  the  undertaking  is  facing  its  “customers”  is  scrutinized.  Of  course,  when  dealing  with  private  
entities  such  as  Google  and  Facebook  etc,  establishing  whether  they  are  undertakings  or  not  may  not  
cause  a  concern.  However,  when  dealing  with  Public  Sector  Bodies,  it  may  cause  problems.  See  the  
Compass-­‐case,  where,  according  to  CJEU,  an  activity  consisting  in  the  maintenance  and  making  available  to  
the  public  of  the  data  thus  collected,  whether  by  a  simple  search  or  by  means  of  the  supply  of  print-­‐outs,  in  
accordance  with  the  applicable  national  legislation,  also  does  not  constitute  an  economic  activity,  since  the  
maintenance  of  a  database  containing  such  data  and  making  that  data  available  to  the  public  are  activities  
which  cannot  be  separated  from  the  activity  of  collection  of  the  data.  ECJ,  Compass-­‐Datenbank,  C-­‐138/11,  
ECLI:EU:C:2012:449  discussed  in  Lundqvist,  (2013)  ,  80  et  seq.    
      38  

the popularity of their respective sites/ecosystems, but is that important in an

upcoming IoT setting, where it will be the amount of and quality of data that creates
market power? How should the relevant market be identified? Are there double or
multisided markets. 104 And is such definition of the relevant market helping the
competition law analysis.105 What about market power based only on the amount and
importance of Data, while the service to obtain the data is for zero price; could such a
106
set up imply market power? Moreover, in reference to the exceptional
circumstance doctrine: accessing Data, is that an exceptional situation, which
requires the application of competition law? Perhaps, the data can be duplicated with
a reasonable effort. Indeed, data may be difficult to monopolize.107 Furthermore, is
there a second (downstream) market that the undertaking is reserving for itself? Is
there an elimination of competition and the prevention of the appearance of a new
product under the case law of Magill, IMS Health and Microsoft? Finally, is the Data
an indispensable input or even an essential facility under the same and similar line of
case law?

Indeed, in the scenario of a potential competitor wanting access to specific, unique,

datasets, indispensable for the business conduct, competition law has an
applicability, but, that scenario is perhaps not so common. Moreover, can
indispensability be considered identified when dealing with a dataset that by its
specific size causes network effects to appear, to the point that the downstream or
neighbouring market is (or will be) monopolized?

This is a controversial issue depending on the definition of Data. Social interaction

websites may tip the market to its favour due to network effects in the form of number
of social connections (i.e. connected friends). But, are the ‘friends’ or ‘connections’
interconnected on a social website Data, and is it Data that can be transferred?
                                                                                                               
104  Evans  and  Noel,  “The  Analysis  of  Mergers  that  Involve  Multisided  Platform  Businesses”,  4(3)  Journal  of  

Competition  Law  &  Economics  (2008),  663-­‐695;  and  Filistrucchi,  Geradin  and  Affeldt,“Market  Definition  
in  Two-­‐Sided  Markets:  Theory  and  Practice”,  10(2)  Journal  of  Competition,  Law  &  Economics,  (2014),  
293-­‐339.  
105  Ibid.  Evans,  Noel  and  Filistrucchi  et  al  agree  that  not  all  digital  markets  are  multisided.  For  an  

interesting  analysis  that  e-­‐platforms  are  not  multisided  markets  see  Newman,  (2014),  3  et  seq.    
106  See  Bundescartellamt,  (2016).  OECD  (2016).  The  German  and  French  Competition  Authorities  (2016),  

7,  14  et  seq.  

107  Sokol  and  Comerford  “Does  Antitrust  Have  a  Role  to  Play  in  Regulating  Big  Data?”,forthcoming  in  

Cambridge  Handbook  of  Antitrust,  Intellectual  Property  and  High  Tech,  (Cambridge  University  Press,  
2017)  http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2723693.  (last  visited  23  November  2016).  
      39  

Possibly, Article 20 of the General Data Protection Regulation does grant users that
right portability of such Data and even personal data generated by the data subjects
activity. 108 Thus, it seems that data subjects may transfer Data, making the dat
subject a market player.

But what about Competition Law? Can access be granted to Data under the
exceptional circumstance doctrine? Notwithstanding the above, the Magill109 “logic”
works well in a data scenario: entities (in the Magill case the publicly owned BBC and
RTE et al), engaging in their primary market or (public) task (producing and
distributing TV programmes), create or collect information, (in the form of TV listings)
that might be copyright protected. They are, under the rules of abuse of dominance,
required to give access to this information (the TV listings), due to its indispensability
and that a refusal would be unjust, to an undertaking that will create a new product
(TV guides). Thus, in the Magill case the appellants were not allowed to reserve for
themselves a secondary market. In a IoT setting the device producers would collect
(create) information from the devices included in the product sold to the consumers.
Likewise, the e-platform providers collect data as a by-product for the service
provided.

The Magill case dealt with unique data in the manner that the TV listings could not be
obtained from any other sources. Also IMS Health dealt with a unique brick structure
(a de facto standard for the industry) developed by IMS and the users in conjunction.
Magill may be used to argue access to certain specific kind of datasets under the
exceptional circumstance doctrine; while, general, user generated, and user
voluntarily provided, data will, however, especially after the introduction of IoT,
perhaps not be indispensible causing the doctrine to be triggered.110 Indeed, in the

                                                                                                               
108  Working  Party  under  Article  29  of  Directive  95/46/EC,  Guidelines  on  the  right  to  data  portability  in  the  

GDPR,  adopted  13  December  2016.  The  guidelines  state  on  p.  9:  “[f]or  example,  a  webmail  service  may  
allow  the  creation  of  a  directory  of  a  data  subject’s  contacts,  friends,  relatives,  family  and  broader  
environment.  Since  these  data  are  relating  to,  and  are  created  by  the  identifiable  individual  that  wishes  to  
exercise  his  right  to  data  portability,  data  controllers  should  transmit  the  entire  directory  of  incoming  and  
outgoing  e-­‐mails  to  the  data  subject.    
109  ECJ,  RTE,  ITP  &  BBC  v.  Commission,  C-­‐241/91  and  C-­‐242/91,  ECLI:EU:C:1995:98    
110  An  argument  frequently  posed  by  the  opponents  of  the  essential  facilities  doctrine  is  that  data  cannot  

be  easily  monopolised:  it  is  non-­‐rival  and,  they  argue,  non-­‐exclusive,  since  there  are  no  contracts  
preventing  users  from  sharing  their  personal  information  with  multiple  companies.  Furthermore,  they  
argue  that  there  are  few  entry  barriers  to  new  platforms,  as  data  is  relative  inexpensive  to  collect,  shorted-­‐
      40  

future certain devices, e.g. cars, refrigerators, mobile phones etc., might be able to
collect same or similar personal data from us as we today provide FB and Google.
Moreover, the data that can potentially be collected by these devices may have a
high quality and be more efficient and effective trying to map the general consumer,
than the information that consumers of social sites on the Internet are currently
voluntarily providing.

The issue whether general competition law (more specifically the exceptional
circumstance doctrine) will be applicable to access general personal data, i.e. the
personal information that people are generating by utilizing the internet, has not yet
been conclusively scrutinized by any competition law court. Nonetheless, it can be
questioned whether a court would find the exception circumstance doctrine
applicable. The doctrine may in a few cases be applicable, it depends on the data set
collected. It will also depend on the actual size and magnitude of the Data collected.
As discussed above, there might be network effects involved in the Data industry,
which may be a reason to make the exceptional circumstance doctrine applicable. If
the Data collected is so vast that it creates a tipping effect making it impossible to
rebuilt a similar data-set, the exceptional doctrine may, possibly, become applicable.
Moreover, the antitrust harm of discriminatory exclusion should be taken into
consideration. Could Google and Facebook be accused of discriminatory refusal to
sell Data vis-à-vis a competitor, when using their Data when selling adds to third
parties.

There are a few competition law cases dealing with Big Data originating from
activities and decisions by EU Member States’ competition authorities. There are a
few French cases, e.g. GDF 111 and EDF 112 concerning discriminatory re-use of

                                                                                                                                                                                                                                                                                                                                                         
lived  and  abundant.  Balto  and  Lane,  “Monopolizing  Water  in  a  Tsunami:  Finding  Sensible  Antitrust  Rules  
for  Big  Data”,  (2016),  http://ssrn.com/abstract=2753249.  (last  visited  14  December  2016),  4  et  seq.  
111  The  French  Competition  Authority  imposed  an  interim  measures  to  GDF,  ordering  that  gas  supplier  to  

grant  its  competitors  an  access  to  some  of  the  data  it  collected  as  a  provider  of  regulated  offers,  in  
particular  consumption  data.  The  aim  of  this  interim  measure  was  to  allow  all  suppliers  to  have  the  same  
level  of  relevant  information  to  make  offers  to  consumers  (no  public  information  or  private  database  
exists  on  households  subscribing  to  gas  contracts).  French  Competition  Authority,  Decision  14-­‐MC-­‐02  of  
09.09.2014.  Due  to  privacy  laws,  the  transmission  of  GDF  data  to  competitors  was  conditional  to  an  
approval  by  consumers.  A  significant  share  of  the  consumers  did  refuse  that  their  data  be  transferred  from  
GDF  to  competing  operators.  The  case  is  discussed  in  the  joint  report  by  the  German  and  French  
Competition  Authorities,  (2016),  20.    
      41  

datasets, where the data sets seem not to have been considered essential or
indispensible for entering the relevant market, but where the data holder used the
data sets in an exclusionary fashion. Access was not granted on fair terms vis-a-vis
potential competitors in a downstream market or in neighbouring businesses.113

A third French case, Cegedim, dealt with medical data used by pharmaceutical
companies to manage their visits to doctors and pharmacies in France. Cegedim
was accused of refusal to sell, on a discriminatory basis, datasets regarding
medical information.114 Cegedim was a leader in the medical database market in
which it has a dominant position, offering both databases and customer
management software to laboratories. It notably produces the OneKey database,
the most widely used database in the industry. Euris, a company that only
produces customer management software but no databases, accused Cegedim
of abusing its dominant position, as Cegedim refused to sell its OneKey database
to laboratories that were using the software marketed by Euris, whereas it agreed
to sell OneKey to laboratories using software developed by other competitors. In
accordance with the Autorité’s precedents as well as European Court of Justice
case law, the Autorité found that access to the OneKey database was not
indispensable to Cegedim’s competitors on the downstream market for customer
management softwares, and thus the OneJey datavbase was not an essential
facility.115

However, the Autorité established that Cegedim maintained a continuous and

unilateral refusal to grant access to OneKey specifically targeting actual or
potential clients of Euris. This difference in treatment of companies otherwise
placed in similar situations constitutes, in the Autorité’s view, a form of
discrimination and Cegedim was unable to provide any objective justification. As
a result, this practice had a seriously harmful effect on Euris (which lost 70% of
its customers between 2008 and 2012) and restricted the laboratories in their
choice of customer management software. Hence the Autorité conluded that
Cegedim had abused it dominant position through exclusionary discrimination.116

Similarly, the Belgian competition authority adopted in 2015 a settlement decision

finding the national lottery body was abusing its dominant position. The lottery was
found to have been using its client database, created under its legal monopoly for
public lottery, when selling its new sports betting products, in a market where they
faced competition, while refusing access to competitors. Indeed, the Belgian lottery

                                                                                                                                                                                                                                                                                                                                                         
112  French  Competition  Authority,  Decision  n°13-­‐D-­‐20  of  17.12.2013,  confirmed  on  that  points  by  the  

court  of  appeal  on  21.05.2015.  

113  A  similar  reasoning  has  also  been  used  in  some  merger  cases.  For  instance,  in  its  EDF-­‐Dalkia  merger  

decision.  European  Commission,  “EDF/Dalkia  en  France”,  COMP/M.7137,  dated  25.06.2014.    

114  French  Competition  Authority,  Decision  n°  14-­‐D-­‐06,  dated  08.07.2014,  relative  à  des  pratiques  

mises  en  œuvre  par  la  société  Cegedim  dans  le  secteur  des  bases  de  données  d’informations  médicales.  
This  decision  has  been  confirmed  on  appeal  but  is  still  pending  in  front  of  the  Cour  de  Cassation  (the  
French  Supreme  Court).  
115  ECN  Brief  (2014).  
116  Ibid.    
      42  

body was re-using unique data sets on competitive markets.117 Possibly, these cases
show a new form of antitrust harm (or “new non-discrimination theory”) whereby a
dominant firm cannot “self-preference” it “own operations over those of competitors”
in a discriminatory way. Indeed, they give some room for the EU Commission finding
a similar antitrust harm in the on-going Google investigation.118

There are also a few Nordic competition authority cases. In Swedish Patent and
Registration Office (PRV)119 from March 2012, PRV started to offer free access to the
Trademark register database to the downstream end-user market, whereas
customers on the upstream wholesale market were offered more detailed data in
different formats (so-called “register lifted data”) for a one-time fee and then a yearly
fee. PRV was accused of marginal squeeze, when selling access to the data in
wholesale market, while giving access for free to consumers on its own website.
PRV, acknowledged being an “undertaking”, was considered dominant in the market
of providing access to the specific Trademark database according to the SCA, while
PRV disputed marginal squeeze and price discrimination charges. In the end, PRV
lowered it fees to marginal cost, and the case was settled.

In The Land Registry, from November 2012, the SCA assessed the way the Cadastre
sold refined information in the land register to commercial private actors. The
complaining re-user purported that the Cadastre was abusing its dominant position
by not giving access to raw data. Instead, the re-user only got access to more refined
data. The SCA did not, based on the facts of the case, find any abuse.120

The above scarce cases show that there is room for competition law in reference to
big data, while access to raw datasets might be difficult under competition law. When

                                                                                                               
117  Platteau,  ‘National  Lottery  settles  abuse  of  dominance  case  with  Competition  Authority’  (29  September  

2015),      
 http://www.internationallawoffice.com/Newsletters/Competition-­‐Antitrust/Belgium/Simmons-­‐
Simmons/National-­‐Lottery-­‐settles-­‐abuse-­‐of-­‐dominance-­‐case-­‐with-­‐Competition-­‐Authority  (last  visited  17  
September  2016)  
118  Cf.  Petit,  Theories  of  Self-­‐­‐Preferencing  Under  Article  102  TFEU:  A  Reply  to  Bo  Vesterdorf  ‘(April  

29,  2015).  Available  at    

SSRN:  https://ssrn.com/abstract=2592253  or  http://dx.doi.org/10.2139/ssrn.2592253  (last  visited  21  
November  2016)  and  Bo  Vesterdorf,  Theories  of  Self-­‐­‐Preferencing  and  Duty  to  Deal  –  Two  Sides  of  the  
Same  Coin,  1(1)  Competition  Law  &  Policy  Debate,(2015),  4.  
119  SCA,  The  Swedish  Patent  and  Registration  Office,  Dnr  470/2011,  from  September  2012.  
120  SCA,  The  Swedish  Land  Registry,  Dnr.  601/2011,  from  November  2012.  
      43  

the Data set is indispensable access may possibly be granted, while there are cases
concerning discriminatory exclusion which might create a duty to deal. Competition
law will still has an important role in reference to Big Data, with possible abuses,
such as marginal squeeze (see PSI related cases), discriminatory access of personal
data, while not being essential data (see French cases and also PSI related cases),
exclusivity arrangements (The EU Commission Google investigation and
collaborations in reference to the cultural sector under the PSI Directive), violation of
data protection rules when obtaining personal data, exploitative abuse (German
Facebook case 121 ), perfect monopoly pricing (eCall), and, moreover, excessive
collection of personal data be perhaps considered an excessive abuse, if the notion
of users are ‘paying’ with personal data when using ‘free’ services on the Internet is
correct.122

Moreover, competition law will still be a useful tool to regulate SEPs when the
standards of the IoT will materialize. It is easy to predict that the future, when the
brick-and-mortar industries will start produce interconnected ‘things’ the notion of
patent war may take a whole new meaning. Indeed, the Huawei case and the issues
it raised may very well be addressed again by the EU Courts.

5.4 Data Protection Rules and Competition Law

The data protection rules, whether they should be used as a benchmark for finding
competition law violation,123 or whether they should be considered, and altered, to
something of a (intellectual) property right (that competition law can ‘trump’) is an
                                                                                                               
121  Bundeskartellamt  initiates  proceeding  against  Facebook  on  suspicion  of  having  abused  its  market  

power  by  infringing  data  protection  rules  

https://www.bundeskartellamt.de/SharedDocs/Meldung/EN/Pressemitteilungen/2016/02_03_2016_Fac
ebook.html.  Date  of  issue,  02.03.2016,  See  also  for  instance,  in  Allianz  Hungária  (2013),  the  ECJ  held  that  
the  impairment  of  objectives  pursued  by  another  set  of  national  rules  could  be  taken  into  account  to  
assess  whether  there  was  a  restriction  of  competition  (in  this  instance,  by  object).  Referring  to  German  
Competition  law,  the  German  Federal  Court  of  Justice  has  stated  that  contract  terms  which  are  
incompatible  with  the  laws  regulating  general  conditions  and  terms  of  trade  might  be  an  abuse  of  a  
dominant  position  if  the  use  of  the  terms  is  based  on  the  company’s  market  dominance”.  ECJ,  Allianz  
Hungária,  C-­‐32/11,  ECLI:EU:C:2013:160.  See  also  German  Federal  Court  of  Justice  (Bundesgerichtshof),  
„VBL-­‐Gegenwert“,  KZR  61/11,  judgment  of  16.11.2013,  para.  68.  
122  Kerber  (2016).  
123  Press  release:  Bundeskartellamt  initiates  proceeding  against  Facebook  on  suspicion  of  having  abused  

its  market  power  by  infringing  data  protection  rules,  

http://www.bundeskartellamt.de/SharedDocs/Meldung/EN/Pressemitteilungen/2016/02_03_2016_Fac
ebook.html.  Date  of  issue,  02.03.2016.  
      44  

issue up for grasp to be solved by researchers and practitioners. Possible, data

protection rules and the interface between data protection rules and competition law
may become topics for a general sector specific regulations in the future.

When re-users or Data brokers are interviewed, the data protection rules are
considered the grand ”show stopper” and the issue for these firms are whether
competition law can be used to override data protection rules. The scenario is
whether a firm holding large amount of personal data obtained through contractual
obligations in consumer users agreements, can be obliged to share this data with a
competitor, or may that firm claim that the data protection rules prevent him from
transferring the data? Would it be different if the firm had previously traded in the
personal data by giving access to other firms for remuneration perhaps within that
firm’s ecosystem? Generally, can competition law trump Data Protection Rules?
From the case law of the EU Commission, it seems difficult. In Asnef-Equifax the
CJEU stated “…any possible issues relating to the sensitivity of personal data are
not, as such, a matter for competition law, they may be resolved on the basis of the
124
relevant provisions governing data protection.” Moreover, in the
Facebook/Whatsapp125 merger case, the Commission states, “[a]ny privacy related
concerns flowing from the increased concentration of data within the control of
Facebook as a result of the Transaction do not fall within the scope of the EU
competition law rules but within the scope of the EU data protection rules.” Finally,
the argument put forward against using Competition Law to trump Data Protection
rules is that reduction in privacy equals reduction in quality, and that is not the same
thing as using competition law to trump an intellectual property right. ‘Quality’ may be
an objective of competition law itself, while upholding property rights is less so.

The European Data Protection Supervisor has in 2014, indicated a shift in policy and
a “more holistic approach to enforcement”, in which a more systematic dialogue is
maintained between competition, consumer and data protection authorities.
However, the collaboration between DG Connect and DG Competition seems rather
unlikely. Nonetheless, given the importance of personal data for the digital economy,

                                                                                                               
124  ECJ,  Asnef-­‐Equifax,  C-­‐238/05  ECLI:EU:C:2006:734    para.  63.  
125  European  Commission,  COMP/M.7217,  Facebook/Whatsapp  dated  03.10.2014,  para.  164.    
      45  

access to personal data under the exceptional circumstance doctrine should in rare
cases be tolerated.

6. Conclusion
 
In this article several aspects regarding interoperability and access in reference to
the Digital Economy has been discussed. Firstly, the IP landscape of the soon to
come Internet of Things is drawn up, discussed and scrutinized, and the current
issue whether non-personal digital data should become a property right in-itself was
analysed. The starting point must be that there is no need for more property rights,
while should we create rights in reference to data, not only non-personal data, should
be encompassed, also personal data must be included in the right given to the data
producer / author. Indeed, there is no natural dichotomy between personal and non-
personal data, and the fundamental rights to privacy and personal data can still be
upheld, even though giving a right to the data collector.

Secondly, in light of the IP landscaped, “pictured” in the first part of the article, the
current standardisation efforts for the Digital Economy are discussed, e.g. what are
the challenges, and how much should be standardized and how lenient should
competition authorities treat pre-standard consortia. The Competition authorities
need to become more refined in their analysis of joint standard-setting in the digital
economy. Indeed, there is an interface in-between consortia, joint standard-setting
and the development of vertical and horizontal ecosystems, that competition
authorities need to take into consideration. A “carte blanc” attitude does not suffice
and can create monopolistic, oligopolistic, if not oligarch structures not only on
individual markets but in whole industries. Thirdly, current and future competition law
issues for the Digital Economy were identified and discussed. Competition law clearly
still has a place and use in the Big Data Industry.126 However, when defining the role
for competition law certain aspects should be taken into consideration: (i) the sources
of data are increasing in number and possibly also in quality; (ii) access seems to be
granted by sector specific regulation, rather than through general competition law.
                                                                                                               
126  Not  alone  of  this  opinion:  Grunes  and  Stucke,  “No  Mistake  About  It:  The  Important  Role  of  Antitrust  in  

the  Era  of  Big  Data”  Antitrust  Source  (2015),  Online;  University  of  Tennessee  Legal  Studies  Research  
Paper  No.  269.  (2015),  available  at  SSRN:  http://ssrn.com/abstract=2600051,  (last  visited  12  November  
2016),  2  et  seq.  
      46  

Nonetheless, competition law still has an important role in reference to standard-

setting, Big Data, with possible abuses, such as, for example, marginal squeeze (PSI
related cases), discriminatory access of personal data, while not being essential data
(French cases, PSI related cases and possible EU Commission investigation into
Google search bias conduct), exclusivity arrangements (Google investigation),
violation of data protection rules when obtaining personal data, exploitative abuse
(German Facebook case), perfect monopoly pricing (eCall) and, even excessive
collection of Data might be an antitrust harm in itself. In fact, perfect monopoly
discrimination can possibly be achieved by utilizing Data on certain markets.
Moreover, collusion (tacit) may be facilitated through utilizing Data and algorithms.127
The integrity of the standard setting procedure must also be held intact, while this
form of creating efficient standards is necessary for establishing markets and
industries. Should dominant firms violate this integrity, such business conduct may
be considered an abuse. Lastly, the unworkable dichotomy between personal data,
non-personal data was criticized.

                                                                                                               
127  Ezrachi  and  Stucke  “Artificial  Intelligence  &  Collusion:  When  Computers  Inhibit  Competition”  Oxford  

Legal  Studies  Research  Paper  No.  18/2015  (2016);  University  of  Tennessee  Legal  Studies  Research  Paper  
No.  267.  (2016);  Available  at  SSRN:  http://ssrn.com/abstract=2591874  or  
http://dx.doi.org/10.2139/ssrn.2591874,(last  visited  12  September  2016).  

View publication stats