EARSC Open data study - Final Report

ABOUT GMES AND DATA : GEESE AND
GOLDEN EGGS
A Study on the Economic Benefits of a Free and Open Data Policy for Sentinel
Satellite Data

Final Report

by Geoff Sawyer and Marc de Vries
Under an assignment from the European Space Agency

EUROPEAN ASSOCIATION OF REMOTE SENSING COMPANIES
Brussels, 6th December 2012

Study on the Economic Benefits of a Free and Open Data Policy for Sentinel Satellite Data

2012 – ABOUT GMES AND DATA : GEESE AND GOLDEN EGGS: Study on the Economic Benefits of a Free and
Open Data Policy for Sentinel Satellite Data
Authors: Geoff Sawyer and Marc de Vries

This work is licensed under a Creative Commons Attribution 3.0 Unported Licence.
This document has been produced with the financial assistance of the European Space Agency and in collaboration with the European
Commission. The views expressed herein can in no way be taken to reflect the official opinion of the European Union and/or the European
Space Agency.

2
ABOUT GMES AND DATA : GEESE AND GOLDEN EGGS – Geoff Sawyer and Marc de Vries

FOREWORD

The context and challenge
The Global Monitoring for Environment and Security (GMES) Programme1, with consequent
investment by the European Union (EU) and Member States of the European Space Agency (ESA),
holds the potential to deliver significant economic benefits to Europe’s fledgling Earth Observation
(EO) services industry and societal benefits to all European citizens. However, differing views of
Member States as to the funding approach not only puts significant pressures on the overall
budget, it also impacts the data policy and associated business model to be adopted: should
governments seek to recover costs of production and processing of EO data by charging re-users,
yielding direct returns, or should these costs be covered from the general budget, seeking longer
term benefits. Put differently, GMES may well be Europe’s goose capable of laying golden eggs. But
how can we ensure a steady sustainable business model: do we take one egg (direct returns from
sales of data) or do we allow the egg to hatch, hoping more golden-egg-laying geese will follow?

The business model paradigm is shifting
Where in the last two decades of the previous century, public sector bodies (PSBs) were being
encouraged to partly self-finance their operations, governments are now increasingly starting to
recognise that charging for the re-use of public sector information (PSI) through a cost-recovery
model does not necessarily lead to the best economic returns. In fact, a nucleus of evidence is
building up that making PSI available at zero or marginal cost can generate significant benefits,
enabling innovation, new business and job creation in the private sector, and increased efficiency in
the public sector.

The need for sector-specific evidence
Interestingly, this evidence is being generated in more ‘traditional PSI domains’ such as the

meteorological and hydrographical sectors and the geographical domain, as well as in the fields of
business registers, and statistical and legal information. However, despite its widely recognized
potentially high returns, no such research has been carried out so far within the GMES sector.
Obviously, facing the policy challenges and the crossroad concerning the data re-use funding
model, such evidence would be of great value to policy makers and stakeholders, allowing them to
take appropriate measures.

The value delivered for the GMES domain
Accordingly, this study addresses this need. It describes the dilemmas faced and provides a
theoretical economic analysis as well as a comprehensive overview of quantitative empirical
research on the economic effects of re-use charging policies for PSI. Subsequently, it transposes
those findings into the GMES domain, assessing the impacts it may have and establishing an
objective framework of reference for fact-based decision making.

The authors
The work on this study has been carried out by Geoff Sawyer and Marc de Vries. Geoff has a strong
background in space technology (engineering) and is the Secretary General of EARSC with
significant experience in the exploitation of EO geo-information services and influencing policies.
Marc is a consultant, with a background in law and economics, who has been working in the field of
re-use of PSI for over 15 years, for governments across Europe and with various European
institutions.

Obviously, the views in this report represent those of the authors.

1
Note that the name of the programme GMES has recently been replaced by the name Copernicus.
3

Table of contents

FOREWORD.................................................................................................................................. 3

EXECUTIVE SUMMARY.................................................................................................................. 6

1 GMES AND OPEN DATA – STATING THE CASE AND INTRODUCTORY REMARKS........................ 8

1.1 GMES: a growth enabler ...................................................................................................................... 8

1.2 PSI re-use and Open Data basics .......................................................................................................... 9

1.3 What data will be available?................................................................................................................ 9

1.4 The PSI re-use financing dilemma ...................................................................................................... 11

1.5 The evidence in other PSI domains is building up .............................................................................. 13

1.6 However, the EO domain remains fairly terra incognita .................................................................... 13

1.7 Europe’s returns could be high .......................................................................................................... 14

1.8 Aims, build up, process, outcome, contributors and dissemination ................................................... 15

2 THEORETICAL FRAMEWORK – THE LAWS OF PSICONOMICS ................................................. 16

2.1 Introduction....................................................................................................................................... 16

2.2 PSI characteristics .............................................................................................................................. 16

2.3 Economic implications ....................................................................................................................... 17

2.4 Wrap up............................................................................................................................................. 20

3 HUH, HUH, WHERE IS THE PROOF? – EMPIRICAL EVIDENCE FROM VARIOUS PSI DOMAINS ... 21

3.1 Introduction....................................................................................................................................... 21

3.2 European comprehensive PSI value studies ....................................................................................... 21

3.3 Overview case studies ....................................................................................................................... 23

3.4 Chain of economic effects of lowered PSI re-use charges .................................................................. 25

3.5 Wrap up............................................................................................................................................. 36

4 WOULD IT WORK WITH US? – TRANSPOSITION IN THE GMES DOMAIN ................................ 37

4.1 Introduction....................................................................................................................................... 37

4.2 EO domain evidence .......................................................................................................................... 37

4.3 Economic effects of free re-use of GMES data – an educated guess................................................... 43

4.4 Wrap up............................................................................................................................................. 45

5 THE CONSEQUENCES – THE WAY AHEAD, HOW TO MAKE IT WORK? .................................... 46

5.1 Introduction....................................................................................................................................... 46

5.2 Further considerations....................................................................................................................... 46

5.3 Policy implications ............................................................................................................................. 47

5.4 Final observations.............................................................................................................................. 51

ANNEX 1 – ABOUT THE AUTHORS ............................................................................................... 53

ANNEX 2 – ABBREVIATIONS AND SEMANTICS ............................................................................. 54

ANNEX 3 – SNAPSHOTS OF THE ‘OLD’ AND ‘NEW’ PSI DIRECTIVE ................................................. 56

ANNEX 4 – STRUCTURED SUMMARIES OF CASE STUDIES ............................................................. 58

ANNEX 5 – NATIONAL DATA PORTALS......................................................................................... 72

ANNEX 6 – LIST OF CONTRIBUTORS............................................................................................. 75

ANNEX 7 – BIBLIOGRAPHY .......................................................................................................... 76

ABOUT GMES AND DATA : GEESE AND GOLDEN EGGS – Geoff Sawyer and Marc de Vries 5

EXECUTIVE SUMMARY

In the 1980s and ‘90s privatisation became a popular policy and treasuries looking for all possible
ways to raise funds found that they were sitting on a pile of golden eggs. Data and information
collected by public bodies in the course of their duties could be sold to help subsidise their costs. In
this way, treasuries could reduce public subventions and in effect raise money. Just as the owner of
the goose that laid the golden eggs became greedy for more so did the public owners of data,
becoming fixated on selling it to earn revenues.

Recent studies show that this policy, far from creating more golden eggs, is in effect, as in the fairy
tale, destroying them. Public sector bodies (PSBs) that sell their data discourage innovation and
creativity, inhibit the creation of new businesses and generally earn less for the government than if
they were to give the data away for free. Indeed, those PSBs that do give their data away create more
overall income for treasuries through increased business and employment taxes. The PSB goose really
can lay golden eggs.

Our study is to examine this policy in the context of GMES. The European Commission has been
working with the European Space Agency (ESA) to establish a system of satellites – the Sentinels – to
provide Earth Observation (EO) imagery from around the world. This data will be used to provide
information services in support of public policy notably for the environment and security. Hence the
acronym GMES – Global Monitoring for Environment and Security.

For a long time it was assumed that the data delivered from GMES would be free to public
policymakers but not for other users. Increasingly, a number of stakeholders, most notably ESA, have
argued that it would be more effective to offer this data free of charge to all users. EARSC supports
this policy as offering the best way to help the emergent EO services industry to develop and indeed

to establish itself as a world leader.

Since many governments, especially at the European level, favour the move to free and open policies
for all types of government-collected data, it would seem logical to apply this policy to GMES data. In
this study we start to bring together the arguments and ideas from the work on PSI re-use and the
issues and needs for exploiting GMES. No links have been made between the world of EO data and
services and the wider world of PSI re-use. It is time to bring them together.

We start the study by looking at the theoretical perspective of how the re-use of PSI is best financed
through general taxation from the State budget, as opposed to charging the users. This is because of
the peculiar economic characteristics surrounding PSI (PSI-conomics), as it is fundamentally different
from classic physical goods produced by companies.

We continue by examining the characteristics of the data and information that will be generated from
GMES before contrasting this with other domains in which PSI is generated. Here we look at cadastral
and cartographic information and data and meteorological services, which indeed are closest to the
world of GMES. We find a number of areas where hard evidence is available on achieved benefits
coming from a change of pricing policy: from cost recovery to marginal or zero cost.

Indeed, empirical research in various PSI domains seems to confirm the theoretical assumptions
reflected in chapter 2. Briefly put, the benefits largely outweigh the costs (and lost incomes of the
PSBs): private sector activities increase, leading to economic growth, more employment and better
services, due to market dynamism. The additional tax returns make up for the extra costs and
incomes foregone by the PSBs. The only issue is that these macro returns only kick in later, requiring
transitory financing arrangements.


Our model for the transition and subsequent benefits is then applied to GMES but only after we have
looked for evidence coming from the EO domain. The best example here is the US Landsat satellite
system that over its 40 years of existence has changed charging policy several times; most recently in
2008, when it moved to a free and open policy. The increase in data use was dramatic with over 100
times more data downloaded in 2011 than in 2007, a result which corresponds with the results in
several of the classical PSI domains mentioned above.

The lessons learned lead us to draw a number of conclusions and make several recommendations to
policy makers on action to be taken. Collectively, these will improve the commercial environment and
enable the EO services industry to grow; hence contributing to much-needed, economic growth in
Europe.

We conclude by showing that evidence coming from the other domains does support the argument
that free re-use of GMES data will offer better exploitation potential. Bringing together these two
different policy areas, both being led from Brussels, strongly supports the stakeholders’ views. A free
and open data policy may finally be seen as “the goose that lays golden eggs”.



1 GMES AND OPEN DATA – STATING THE CASE AND INTRODUCTORY
REMARKS

1.1 GMES: a growth enabler

Not a goose but laying the basis for future growth
Global Monitoring for Environment and Security (GMES) is a publically funded programme with the
goal to ensure that European policy makers have access to globally derived geo-information products
necessary to inform policy decisions. At its heart are a number of Earth Observation (EO) satellites
(Sentinels) that provide data which can be turned into information used to inform on environment
and climate change, emergency management, natural resource management, security of the citizen
and many other public tasks.

As the European Commission has noted: “At a time when command and appropriate use of
information has important geo-strategic implications, Europe needs to have available a capacity
which allows it independently to evaluate its policy responses in a reliable and timely manner. A
comprehensive Earth observing system, using space-borne and in situ techniques (land, air and sea)
through well defined, operational services, is key to ensuring the implementation and monitoring of
environmental and security policies in the context of sustainable development.”2

But the information generated by the GMES system, and especially the EO data, will be valuable not
just for the European public sector but also for many industries and public organisations outside of
Europe that require spatial geo-information. Hence, a free and open data policy also has the
potential to stimulate the European EO services industry and help it develop into a global leader.

Killing the goose
Nevertheless, this view is not shared by everyone, with some proposing that the commercial sector
be charged for data and information. A similar cost-recovery model policy was first advanced in the
1970s and ‘80s by treasuries as a means to recover costs incurred in the collection of information by
some government bodies, e.g. meteorological agencies, cadastral authorities. As a result, this
information has been sold at often rather high prices to a limited number of users.

Recently, this policy is under challenge and, in some countries, has been at least partially changed.
Evidence is growing that by making the data freely available, greater economic benefit can be
realised. Companies downloading the information are able to innovate, generating new products and
services, hence leading to wealth creation. Those following the alternative cost-recovery model are
in fact killing the goose that can lay golden eggs.

GMES and Earth Observation (EO)
Throughout the report we refer to GMES and to EO. In this context we use the term GMES where it
applies directly to the programme of that name. It may also be used to refer to the overall system
that will result. Earth observation or EO is a more general term and consequently is used to refer to
the wider activities associated with satellites, products and services that are not themselves covered
by GMES.

2
SEC (2005) 1432, Global Monitoring for Environment and Security (GMES): From Concept to Reality;
Communication from the Commission to the Council and European Parliament.


1.2 PSI re-use and Open Data basics

PSI Directive acknowledges PSI as a driver of economic growth
Carrying out its public tasks, the public sector collects, creates, produces and disseminates a wide
variety of information, such as legal and administrative information, business and economic data, or
geographic and meteorological information. With the advent of the digital age in the 1990s, people
started to recognize that quite often this public sector information (PSI) has a ‘second value’, where
it constitutes a valuable raw material that can be re-used by third parties in added-value information
products and services.

Against this background the PSI Directive was adopted in 2003, seeking to stimulate further the
development of a European market for PSI-based services, to fortify competition in the internal
market and to address divergence as to re-use rules between Member States. In essence, the
Directive imposes obligations on Member States’ public sector bodies (PSBs) once they decide to
allow re-use of their PSI and provides corresponding rights to re-users thereof. These obligations
include transparency of conditions of re-use and tariffs and requirements to apply equal conditions
to comparable types of re-use, including their own commercial re-use and a ban on exclusive
agreements.

Second generation PSI re-use rule underway
After a fairly slow start, the Directive – most prominently represented by the fast growing ‘Open
Data community’ – became increasingly popular, and in December 2011, EU Commissioner
Ms Neelie Kroes launched a proposal to update the Directive in order to align it with the revolution in
technologies we have been witnessing. Briefly put, the new proposal tightens the obligations of PSBs,
particularly on the issue of fair pricing and transparency. The European Parliament and the Council
are now reviewing the proposal in a co-decision procedure. Assuming that the proposal will be

adopted in the spring of 2013, Member States will have a transposition period of 18 months,
meaning that the new rules will have to be complied with by 2015.3

1.3 What data will be available?

Characteristics of the GMES value chain
The system that is being constructed under the GMES programme will generate several different
types of data. The value chain for EO geo-information services at its simplest level is shown in Figure
1-1, where satellite imagery is downloaded by a ground station and made available to value-added
processing organisations. The satellite data is processed into information products through a process
of value-added (sometimes referred to as downstream processing), which normally involves
assimilation with other datasets (other satellite imagery or ground-based in situ data), before being
made available to users. In the case of GMES, the users are primarily in the public sector.

Therefore, in this value chain, we have essentially three different types of data: satellite imagery, in
situ data, and processed information, collectively referred to as GMES data and information.

3
For those seeking more details on the 2003 PSI Directive and the 2011 proposal, Annex 3 holds snapshots of both.


Figure 1-1: GMES value chain

In this chain, data from satellites or from in situ measurements are processed into information by a
value-adding process. This information is specific to the GMES services mentioned earlier.

a) Satellite imagery
The satellite data to be used for GMES comes primarily from the Sentinel satellites being developed
by ESA. They are satellites funded and operated by the public sector. However, data from other
satellites (nominated as contributing missions) will also be used.

Sentinels provide medium and low resolution imagery (10m or greater), which in itself has a low

commercial value. The satellites are designed to give regular, moderate resolution imagery of a large
part of the Earth’s surface. The major use is for public tasks since the sensors measure parameters
that have more interest for science and long-term climate monitoring. The data has characteristics
suited to specific uses, i.e. atmosphere, ocean and land monitoring. Combining data in time series
and with other observations can greatly increase its value.

GMES Sentinel data is fully paid for and owned by the public sector and hence is a core PSI.

Contributing missions will mainly provide high resolution optical or radar imagery (generally better
than 5m)4. They are mainly owned and operated by private companies (commercial operators),
acknowledging that the private investment has often been made alongside some government
investment. The imagery is higher in resolution than that produced by the Sentinels (i.e. around 1m
in the cases of Pleides, Cosmo, TerraSar-X and around 5m in the case of RapidEye, DMCI). The data
has characteristics best suited to emergency management and surveillance activities.

Contributing mission data is owned by the satellite owners and licensed to the public sector. As such
the degree to which this data may be PSI will depend on shared rights determined by the terms of
the licensing conditions.

b) In situ data
This is data collected by PSBs (agencies) generally in pursuit of their public task. It comes from
ground-based, ocean-based and atmospheric sensors and will be used with the satellite imagery to
improve accuracy and validate the final information products. In some cases, the in situ data may

4
Note: we are generally using 5m as the boundary between low/medium resolution imagery and high
resolution. We are not distinguishing between high and very high resolution as is the case in many studies as it
does not bear relevance. The highest resolution GMES Sentinel data is 10m.


belong to specific end users (e.g. an NGO in the case of emergency situations) and may or may not be
shared with others. Where the data is from a PSB then we presume it will be classified as PSI and
hence should be made available for re-use.

c) Processed information
Processed information is that which shall be made available through the GMES services. Clearly it is
derived from combinations of satellite imagery and in situ data. The GMES services will be paid for by
the public sector out of an EU budget; it fits the criteria to qualify as PSI.

The parallel with PSI re-use is evident: EO data (coming from satellites) and the information (coming
from processing the satellite data in combination with other data) being produced under the public
task can also be used for other purposes, providing a wealth of material to be exploited in new
domains. It will likely be a strong lever to help develop a European services industry, competitive on
a global stage. Put differently: GMES will produce high-end PSI.

Clearance of rights essential
But what about the contributing missions? The mere act of (investing in) collecting and adding value
to GMES data creates intellectual property rights (IPRs), in particular copyrights and database rights.
Obviously, where appropriate, free-flow and re-use of GMES data will require clearance of these
IPRs. Where these rights pertain to European PSBs (like ESA or the EC or even Member States’ PSBs),
re-use can simply be allowed by waiving or not exercising these rights. However, where the IPRs are
held by private sector parties or individual countries, clearance of these rights will be needed to
allow re-users to legitimately re-use the data, incorporating them in to services and products.

The EO data coming from contributing missions will be procured under licence, negotiated by the EC,
leaving certain rights with the satellite operators, regardless of whether they are commercial
companies or nation states. The extent of the residual rights depends entirely on what is negotiated

with their owners, which depends in turn on the nature of payment – whether in cash or in kind. If in
kind, this may be by negotiating reciprocal rights to a third party’s mission. If in cash, as in a
commercial transaction, the extent of the rights will depend on the amount paid. Ideally, the
contract would give the EC rights that are equivalent to the GMES Sentinel data – which would mean
free to everyone. Clearly, a commercial company will need a large contract to offer such a licence
since they preclude significant third-party sales. Hence, the precise rights to be negotiated will
depend on commercial evaluation and on the budget available.

1.4 The PSI re-use financing dilemma

PSI re-use = second use
As mentioned, PSI re-use is capturing the hearts and minds of policy makers across Europe.
Increasingly, they recognize that information produced within the context of the public task, often
has a ‘second’ value: a re-use value. This re-use value is to be cashed in by re-users: mostly
businesses and increasingly non-commercial users, Open Data activists in particular, who seek to
avail themselves of this rich resource, creating new products and services.5

Charging for re-use: two camps
One of the central themes within the discussions taking place concerns the effects of charging for re-
use by the PSBs holding the PSI.

5
Throughout this study numerous key terms are used. Grasping their exact meaning from the outset is
essential. Annex 2 therefore holds a list addressing these semantic issues.


PUBLIC SECTOR
Public Sector Public sector
State Body carrying out User
Treasury fee user
budget public task

creates
Registration+ Public Delivery to
updating sector users
information
U
s
e
r
Non-public
fee sector re-user
PRIVATE SECTOR + CITIZENS

Figure 1-2: The two financing options: user fee or State budget financing through the treasury

Figure 1-2 captures this picture. Briefly, there are essentially two (non-exclusive) options available.6

(1) the data holder charges the user
In this option the data holder relies on charging users of the data. This means that the data holder

determines a price, or a range of prices, for the various quantities and formats of data delivered.
Such prices can differ between types of users, particularly if there are users from the public sector
using the PSI within the context of the public task.7 Parties outside the government seeking to re-
use the data, however, are often charged, exceeding the marginal costs for distribution (= cost-
recovery pricing). Regardless of the user (public or private), this financing model hooks on to
output of the data holder, allocating the costs over those who need or want to avail themselves of
the data (either under a public task obligation or for re-use in their business processes).

(2) the data holder is paid from the State budget
Here, the option is to facilitate direct financing from the State budget. As such there is no direct
relation between the budget and the input and output of the data holder, but rather on a regular
(yearly) assessment of cash required to run the operation. The costs are not passed on to the
entities registering or updating registrations or to the users of the data, but are divided among all
taxpayers. No charge is imposed for the use or re-use of the data – neither on public users nor on
private sector re-users.

These two options correspond with two opposing views. On the one hand there is a growing crowd,
often referred to as the ‘Open Data movement’, advocating that society will gain maximum benefit if
companies are given free and open re-use rights to data. They argue that such re-use policy will lead
to maximum economic benefit, as this will provide incentives at the least cost and the maximum
opportunity to develop new markets, whereby the increased tax returns and efficiency will outweigh

6
For some categories of PSI there is a third option: charging for the registration and/or updating. Many
business registrations (also) rely on this income stream, charging new companies for registration and charging
yearly fees for an extension of the registration. Accordingly, such PSBs face a trilemma rather than a dilemma.
7
Increasingly, it is recognized that internal charging within the government does not make much sense. In the
context of this study, we will not further discuss this option.


the losses incurred by dropping user charges. On the other hand, there are those who favour cost-
recovery models whereby the PSB is allowed to charge for re-use of its PSI, turning it into an income-
generating source, decreasing the general tax burden, passing costs on to those who use the data.
They also question the sustainability of a scheme providing PSI at no cost or marginal prices when the
costs of creating and maintaining quality PSI can be substantial (requiring additional public funding).

Charging options represent an economic dilemma
So in essence, the economic dilemma is how to finance operations: by charging users or by
recovering the costs from the general State budget? The good thing about this dilemma is that, to a
large extent, the decision can be largely based on economic laws and facts and figures, where the
yardstick should be: will the economic effects (costs and benefits) of charging marginal or zero costs
for re-use of data outweigh the economic effects (costs and benefits) of cost-recovery models.

Current EU rules allow for cost recovery, but favour marginal cost pricing
At present, although fostering marginal or zero cost pricing regimes for PSI re-use, the European
regulatory framework – through article 6 of the PSI Directive – gives PSBs the right to charge for the
re-use of their PSI, thereby generating an income that should not exceed the cost of collection,
production, reproduction and dissemination, together with a reasonable return on investment. The
more general competition law framework under the Treaty on the Functioning of the European
Union (TFEU) appears not to limit this position further (with the exception that a PSB's charging
conduct could be regarded as a clear abuse of a position of dominance).

Relevance for EO data high
Earth Observation (EO) data is potentially a highly rewarding PSI domain. Obviously, the dilemma
also emerges in this domain, particularly in the context of on-going discussions between Member
States in the GMES Programme. Differing views of Member States as to the funding approach not
only puts significant pressure on the overall budget, but it also impacts the data policy and

associated business model to be adopted. We shall deal with this further in chapter 4.

1.5 The evidence in other PSI domains is building up

Surge in PSI re-use research
Research into economic effects of charging by PSBs for re-use of their data has recently enjoyed a
surge in popularity. Where previously only a few authors were concerned with this issue, over the
last few years a steady stream of reports and studies has sprung from the academic world, policy
makers and re-users.

Where most of these studies conclude that there is a clear business case for zero or marginal
charging, there is also a growing body of evidence which strongly suggests that taking away such
barriers for re-use has beneficial impacts not only on the downstream market (i.e. on economic
growth, employment and innovation), but also in the public sector itself, due to observed increases
in efficiency and effectiveness.

1.6 However, the EO domain remains fairly terra incognita

Evidence in the relevant domain is lacking
Previous studies have not made the link between GMES and the case for free and open data and as a
result there is little evidence in the EO domain to support the case. The strongest evidence comes
from the US Landsat programme that has made the transition from a cost-recovery model in the
1980s and ‘90s to a free data policy that was only fully implemented in 2008. This change in policy
provides some evidence, which we shall examine in chapter 4, but no real market data yet exists on
which to base concrete, financial analysis.


This has not prevented several stakeholders from making the argument in support of a free and open
data policy. The ESA Council first proposed this in 2010, and during 2011, EARSC published three
position papers on the subject of GMES: (1) Exploiting GMES Operational Services (March 2011), (2)
The Threat to GMES (July 2011) and (3) GMES Data and Information Policy (October 2011) providing
the EO services industry perspective and examining the conditions whereby the private sector –
downstream services industries – could deliver further returns, over and above the public sector
economic benefits, on the public investments.

In other words, there is action on the ground suggesting that, in the field of GMES, there is a business
case supporting low or zero cost re-use models. However, so far, no focused in-depth research has
been carried out. This study seeks to make a first and important step in addressing this gap.

This being said, it is important to note that in this study we are not trying to bring evidence to justify
the GMES programme on socio-economic grounds. This has been covered in previous work:

− PricewaterhouseCoopers report for ESA on the socio-economic benefits analysis of GMES
published in 20068.
− Booz & Co report for the EC on cost-benefit analysis for GMES published in September 20119
− The socio-economic benefits of GMES10 published by the European Space Policy Institute.
− In addition, the Ecorys report11 on the competitiveness of the downstream sector (2008) –
heavily relying on earlier studies made by Vega12, 13 and Euroconsult14 – came up with some
data policy considerations.

This work collectively provides overwhelming evidence of the value that GMES can deliver for the
public sector. The goal of our work goes beyond this to the additional value that may be created in
the private sector. Furthermore, whilst our primary objective is to look at the potential value in the

data / imagery coming from the GMES Sentinels, much of the logic will also apply to the information
generated by the GMES Services.

1.7 Europe’s returns could be high

Opportunity knocks?
The GMES programme, born at Baveno in 1998 as a means to develop and sustain European capacity
in the provision of satellite-derived information for the environment and security, now plans to
launch its first dedicated satellites (the Sentinels) in 2013. The system is expected to be fully
operational by 2016. With so much value captured in the massive amount of data to be collected,
there is an urgent need for an appropriate data policy that supports not only the GMES objectives,
but also looks at the potential economic side benefits (positive externalities) that will result from
taking the right decisions, when addressing the policy dilemma referred to above.

Clearly, Europe’s fledgling EO services industry has a leading role to play in the exploitation of the
public investment through selling services built upon GMES data and information. Earth Observation
is inherently global in nature and hence there is strong potential for exploitation through exports as

8
Socio-economic benefits of GMES, PricewaterhouseCoopers, October 2006.
9
Cost-benefit analysis of GMES, Booz & Co, September 2011.
10
The socio-economic benefits of GMES, Giannopapa, European Space Policy Institute, November 2011.
11
Study on the Competitiveness of the GMES Downstream Sector, Entr.06.054; Ecorys, November 2008.
12
The State and Health of the European and Canadian EO Service Industry; Vega and Booz Allen & Hamilton,
2004.
13
The State and Health of the European and Canadian EO Services Industry in 2006; Vega, 2008.
14
Assessment of the Downstream, Value-Adding Sectors of Space-Based Applications; Euroconsult 2007.


well as new customers in Europe. In line with the Lisbon treaty, information services can provide
strong possibilities for growth and economic benefit. Giving the industry free and open access to
GMES data and information should be the most effective way to realise this potential. So setting the
appropriate policies, not just for GMES but in areas where geo-information provides for the
implementation of policy decisions (for example agriculture), is key if the potential is to be realised.

Timing is perfect
This is an excellent moment to demonstrate the benefits that can be gained from a free and open
data policy for GMES Sentinels. The question of GMES financing is on the table and will be decided in
the next few months – most likely in early 2013. Many companies are looking forward to the satellite
data becoming available after the first Sentinel is launched next year and can be encouraged to
commit resources by a strong message. In addition, the EC is currently preparing its data policy
proposals for approval in the next few months.

Furthermore, addressing this issue is in perfect harmony with the EC’s recently launched action to
amend the 2003 PSI Directive15, which, as a rule, will likely impose a maximum charging level
(marginal costs for distribution only). This proposal for a new PSI Directive is now being discussed in
the Council and Parliament, and results are expected in the fall of this year. (For those not quite
familiar with the legal framework on PSI re-use, snapshots of the current PSI Directive and the
proposal for a new directive are attached as Annex 3.)

1.8 Aims, build up, process, outcome, contributors and dissemination

Clear objectives, clear target groups
In this context and against this dynamic background, this study aims to provide evidence of the
potential benefits of a free and open data policy, allowing policy makers at the European and

national levels to take well-founded and sound decisions on re-use policies, based on transparent,
objective and economic criteria.

Thus, the study:
− provides a theoretical economic analysis of the economic laws surrounding PSI re-use: ‘PSI-
conomics’ (chapter 2), and
− gives a comprehensive overview of existing evidence in other (non-GMES) PSI domains:
quantitative empirical research on the economic effects of re-use charging policies for PSI
(chapter 3), then, to the extent possible,
− transposes those findings into the GMES domain, assessing the impacts it may have and
establishing an objective framework of reference for fact-based decision making (chapter 4),
and finally,
− concludes with a set of coherent recommendations (chapter 5).

Appropriate level of stakeholder input and quality control
Maximizing the practical value of this study, we have sought constant interaction with stakeholders.
This included presentations of the contours and first findings of the study during the EARSC Annual
Meeting in Brussels on 6 July 2012. Subsequently, regular sharing of interim results has taken place
in close collaboration with ESA: progress meetings were held in Brussels (2nd August 2012) and Rome
(13th September 2012). Then an Expert Workshop was held in Brussels on 7th November 2012,
hosted by the European Commission DG Enterprise, with experts invited from around Europe.

A list of all persons that contributed to the study is attached as Annex 6.

15
Directive 2003/98/EC of the European Parliament and of the Council of 17 November 2003 on the re-use of
public sector information, OJ L 345, pp. 90–96.


2 THEORETICAL FRAMEWORK – THE LAWS OF PSICONOMICS

2.1 Introduction

In the previous chapter we have seen that there are basically two options to finance PSI re-use: (1)
charging the users of the data and (2) using taxpayers’ money from the State budget. In the following
paragraphs we shall seek economic points of reference to decide on what to base the decision.
Accordingly, these points of reference centre the underlying assumptions of economic theory:
productive resources are scarce and have alternative uses, forcing economic actors (consumers,
companies and governments) to choose between these alternatives in a rational manner to maximise
benefits.

2.2 PSI characteristics

Old school economics does not apply
PSI is an information “good”, a non-physical, electronic “good” produced by the public sector and
consumed both within and outside the public sector. These features cause ‘classic’ economic laws to
work very differently, in comparison with a physical product produced and sold by a company. The
differences in the economic ‘behaviour’ of PSI are caused by two fundamental characteristics: non-
rivalry and non-excludability.

PSI is non-rivalrous and non-excludable
PSI is non-rivalrous, meaning that it can be consumed by one person without detracting from the
consumption benefit enjoyed by any other consumer. In other words, use of PSI by one is not to the
detriment of the others. Furthermore, information products are, in principle, non-excludable: it is

difficult to include one user while excluding others. Where providers of information try to do so, they
tend to rely on limiting access, exercising their intellectual property rights and charging. Obviously,
this comes at a cost, since information, by its very nature, seeks to be disseminated. Preventing and
controlling this dissemination will consume economic resources.

If a good is non-rivalrous and non-excludable, the private sector is unlikely to produce it in adequate
amounts. Herein lies the rationalization for public provision of these goods, especially where they
also have recognized social value and positive externalities. Public broadcasting, national defence,
and lighthouses are perfect examples of such public goods.

Implications for financing options
Justification for charging of public goods is usually based on the so-called benefit principle: public
services are charged for in a way that those who benefit (more) pay (more). This principle
encourages those who benefit to appreciate that there are resource costs involved and it decreases
the taxation burden on those who do not benefit (like levying charges for acquiring building permits
and passports). It takes the detrimental effects into the equation: some users lose out as they do not
want to pay the price. So in essence the dilemma is – as addressed in chapter 1 – what is more
important: equity considerations, which require the beneficiary to pay, or efficiency considerations,
where all those taxed pay and where the under-utilization of the service needs to be weighed against
the distortions associated with general taxation. These characteristics have significant implications
when considering the financing model.

16

2.3 Economic implications

The invisible hand does not work
In economic theory the allocation role of prices requires that they be set equal to marginal cost to
achieve economic efficiency: a producer of a good will continue to increase production as long as the
increase in returns exceeds the costs for producing an extra unit. Accordingly, prices move to the
level where they equal the marginal costs and maximize the efficiency.

Interestingly, a key feature of electronic information goods in the internet age is that for any level of
production, the cost of providing it to an additional consumer – the marginal production costs – is
(close to) zero. This implies that charging for them is never economically efficient, since some users
are prevented from enjoying the good even though their consumption of the good would not
generate any extra (marginal) cost. Accordingly, goods are not purchased and, as a result, they do
not provide benefits to society.

Pricing lacks factual basis
According to neoclassical economists, by allowing prices to move freely the supply of any given
commodity will ultimately match demand. Accordingly, prices determine (a) what goods are to be
produced and in what quantities, (b) how the goods are to be produced and (c) who will get the
goods.

Following these assumptions, economic actors would respond to lowered demand by lowering prices
and production (and vice versa) and thus increase efficiency. However, in real life, publicly produced
public goods and services are not subject to such market forces. Thus, user charges cannot truly
reflect demand, since the governmental offer cannot be refused.

Alternatively, if market-oriented pricing does not work, one can also rely on cost-benefit pricing.
Cost-benefit pricing is based on the analysis whereby the equivalent money value of the benefits and
costs to the community of a certain project are compared in order to ascertain whether it is
economically worthwhile to pursue. However, one of the problems is that although the computation
of many components of benefits and costs is intuitively obvious, there are others for which intuition
fails to suggest methods of measurement. Likewise, assigning costs and benefits for the production
of public goods is burdensome and largely arbitrary.

As a consequence, therefore, if the information for setting efficient charges is not available, the
yardstick for taking well-founded decisions is not available, which makes it practically impossible to
assess whether the ‘right’ decision has been taken.

Cost structure calls for mass production
Producing information features a cost structure radically different from that for physical goods. First,
information is costly to produce and cheap to reproduce, featuring high ‘first-copy costs’ but low
‘second-copy costs’, since the variable cost of the production and distribution of an extra unit, in
particular in a digital and networked environment, does not increase, even if it concerns large
amounts. So, generally speaking, the cost per extra unit sold is fairly stable irrespective of whether
two copies or two billion copies are made.

Further, the (large) fixed costs of information production are regarded as so-called ‘sunk costs’. Sunk
costs are costs resulting from past decisions and cannot be recovered if production is halted.
Accordingly, they can be ignored when making new decisions about production, since they no longer
represent meaningful alternatives.

17

Applying these characteristics to the PSI re-use financing dilemma we can conclude that: (a) where
the PSI has been produced already under the public task PSI, the associated costs are to be regarded
as sunk costs, (b) second copies and distribution of PSI hardly lead to extra costs and (c) charging for
PSI re-use lacks factual basis by definition. Accordingly, economists would argue that the greatest
social benefit is achieved when such goods are priced at marginal cost because there is little or no
incremental cost in disseminating information widely.

Charging comes at a cost
First, charging users entails transaction costs to the seller and to the buyer, since it requires
recording and collecting, and recording and payment, respectively. Second, charging only works if the
free flow of information is disabled, as otherwise buyers will pass on the information to subsequent
buyers, underselling the original provider. Where providers of information try to do so, they tend to
rely on limiting access and/or exercising intellectual property rights (copyrights and database rights
in particular). As these barriers only work if rights are enforced, this requires monitoring and policing,
creating an extra overhead. Third, setting up and running such charging arrangement represent
opportunity costs where those involved might be able to use the resources consumed in another
way, yielding other benefits. Simply put: charging is costly.

Marginal cost and zero cost pricing makes sense
Figure 2-1 visualizes the supply and demand curves and the cost structure in a monopolistic market
for PSI goods. ‘Normal’ (non-governmental) monopolists would seek “point a” (quantity Qm, against
price Pm) on the demand curve, maximizing profits, but causing a deadweight loss (equal to the size
of the triangle abc and representing the allocative inefficiency that occurs since the equilibrium
where demand meets supply is not matched). Governments, however, will likely seek “point d”
Symbol(s) explanation
D Demand curve

MC Marginal costs curve
AC Average costs curve
MR Marginal revenue curve
Qm, Pm Price and quantity a monopolistic supplier would
normally offer, creating a deadweight loss of the size
of triangle abc
Qr, Pr Price and quantity a governmental monopolistic
supplier would offer, in search of breaking even,
reducing deadweight loss of the size of triangle cde
Qc, Pc Price and quantity a governmental monopolistic
supplier would offer, eliminating the deadweight loss
all together, but requiring some cost recovery from
the State budget (taxes)
Qo, Po Price and quantity a governmental monopolistic
supplier offers when it considers the economic
benefits from zero pricing higher than from charging
any price, fully relying on the State budget

Figure 2-1: Supply and demand and cost structure in a monopolistic market for governmental
information goods

(quantity Qr, against price Pr, breaking even, reducing the deadweight loss to the size of triangle cde)
or even “point c” (quantity Qc, against price Pc, where the marginal costs are equal to the average
costs, eliminating the deadweight loss all together, but no longer fully covering the average costs,
requiring some tax money to subsidize the losses). Where the marginal transaction costs (sending

18

bills, monitoring payment, etc.) are relatively high, governments may decide to drop charging all
together (Po), moving to zero costs charging and covering all costs from general taxes.

Collateral damage is to be assessed
Where the consequence of following the efficiency principle is in fact that public goods are priced at
the level of marginal costs, which in a digital environment equals zero, obviously such price setting
would heavily affect the position of competitors in the upstream market.

There are certain PSI domains in which we witness this, for instance in the field of geographic
information, where Open Street Map and Google Maps offer alternatives for governmental
topographic information. Arguably, if the PSB providing the PSI is under the legal obligation to under-
price, this will have detrimental effects on companies competing with the PSB in the upstream
market. Interestingly, however, upstream services such as Google Maps (also) come for free or at
least are based on business models that do not rely on selling the content.

In any case, in the field of production and maintenance of EO data this should be a point of
consideration, where there may be detrimental effects to upstream competitors. We noted earlier
the dual player (public – private) particularity of the sector. The drive to increase private investment
in space infrastructure has led European governments to seek commercial capital to be invested
alongside public funds. These so called public-private partnerships (PPPs) mean that commercial
returns are sought by the private sector irrespective of public involvement.

The systems concerned generally offer data which is different and complementary to that from
GMES. Nevertheless, in the margins there is some overlap where free data may displace that from
commercial sources hence conflicting with the interests of the downstream sector to have free and
open access to GMES data. Careful attention must then be given to ensure that a PSI re-use data

policy does not undermine the commercial investment made in the upstream data provision.

Downstream effects of third-party re-use
Increasingly, governments are deciding to ban internal charging (on the grounds described above),
but appear to be more hesitant about not charging for re-use outside the public sector. Applying
such user charges, governments tend to restrict distribution by relying on intellectual property rights.
Since the data is (partly) disallowed from flowing freely downstream, or rather from networking,
effects are limited.

Recent studies acknowledge that re-use, in particular for the low-end of the market, features
enormous socio-economic benefits through their so-called long-term effects, representing the
consumer surplus ready to be cashed in on. (In chapter 4 we shall look more at this evidence.)

The link with EO data is evident, where this data has significant potential to serve as a basis for
private sector players – in particular those on the low-end market, able and willing to open up
dormant consumer surplus – to use the data for new services and apps. Charging and (consequently)
applying intellectual property rights will likely hamper such development and comes at a social cost
where extra tax returns and employment are missed out on.

Liberalized re-use regime effects on own exploitation are marginal
Obviously, when a PSB lowers its charges to the level of marginal costs this will impact the revenue
stream generated by its own exploitation. This particularly plays a role when a data holder is adding
value to the data itself in the downstream market, in competition with private sector players. Quite
often such a PSB, exploiting PSI in this way, has become reliant on this income and it ends up cross-
subsidizing its commercial activities with its public task activities. In such cases, the PSB will try to

19

charge high prices for its raw data (where it is a monopoly), allowing it to cut prices in the added-
value market (where it is competing with private actors). From an economic perspective, such a
setup is inefficient.

Breaking this deadlock will require transitory funding since moving to lowered charges will affect the
cash position of the PSB as returns will only arrive later. Furthermore, these will accrue mostly
outside its accounts since they will be received directly by the tax authorities, through increased tax
returns.

2.4 Wrap up

From a purely theoretical economic perspective, where we take the macroeconomic and social
welfare of society as a whole as the starting point, then financing of re-use of PSI from general
taxation appears to be most beneficial, as it distributes benefits across society efficiently. Cost
recovery does not: it raises extra costs and disallows the benefits from hatching and flowing freely
across the value chain.

Obviously, the question is: does this work in practice? For that we need to look into empirical
evidence supporting this theoretical analysis, particularly in those PSI domains close to EO data.


20

3 HUH, HUH, WHERE IS THE PROOF? – EMPIRICAL EVIDENCE FROM
VARIOUS PSI DOMAINS

3.1 Introduction

In the previous chapter we have seen that from a purely theoretical perspective charging marginal or
zero costs for PSI re-use appears to be macro-economically more efficient than cost-recovery
models. However, does this work in practice? Empirical research into economic effects of charging
for re-use of PSI has enjoyed a surge in popularity, in particular in the last four years, where a steady
stream of reports and studies has sprung from the academic world, policy makers and re-users.16

Quantitative market measurements and impact case studies
In this chapter we will examine this empirical research, listing and highlighting key outcomes, with a
view to transposing these findings into the EO data domain. We will start off with studies that have
tried to quantify the European PSI market, i.e. the so-called PIRA, MEPSIR, MICUS and Vickery studies
(paragraph 3.2). We will then move down to the meso and micro level covering sectoral studies and
case studies, first providing a schematic overview (paragraph 3.3) and then providing concrete
figures on economic effects resulting from PSBs moving from a cost recovery towards marginal cost,
or even zero cost, PSI re-use charging model (paragraph 3.4). Some of these case studies are part of
larger studies, like for instance the POPSIS study that covers 21 case studies (five of which we will be
referring to). Paragraph 3.5 holds the wrap up.

Focus on monetary data
Looking into the cases, we will mainly focus on the monetary effects that accrue in the value chain:
the costs and benefits. That being said, it is important to realize that some effects, benefits in

particular, simply cannot be monitored and/or monitized. So, how can we make a reliable
assessment of the value of improved data quality resulting from (more) users’ feedback or increased
cooperation and sharing of data between PSBs due to standardisation and enhanced
interoperability? And what is the monetary equivalent of being able to see at the local level whether
it will rain in the coming hours, allowing outdoor workers to plan their work more prudently and
preventing people from catching the flu? So, although we shall (and can) not measure it, we should
not forget that these ‘socio-economic benefits’ are there (and probably in immense volumes).

3.2 European comprehensive PSI value studies

Before moving to the case studies – detailing the effects of lowered re-use prices – we first provide a
snapshot of the efforts to assess the overall value of PSI within Europe.

PIRA
Although the notion that there is significant value in PSI has been gaining weight during the last
decade of the previous century, best mirrored by the 1998 EC Green Paper on PSI17, it was not until
2000 that the first serious attempt was made to somehow connect a figure to this value: the PIRA
report.18

16
Annex 7 holds an extensive overview of sources on economic aspects of PSI re-use.
17
Green Paper on “Public Sector Information: a key resource for Europe” (COM(1998) 585), adopted by the
Commission on the 20th of January 1999.
http://ec.europa.eu/information_society/policy/psi/archives/news/index_en.htm
18
Commission of the European Communities, 30 October 2000. Commercial exploitation of Europe’s public
sector information: Final Report for the European Commission Directorate General for the Information Society.
Pira International. ftp://ftp.cordis.lu/econtent/docs/2000-1558.pdf
21

PIRA estimated that the European market for PSI in 2001 (15 Member States) amounted to an
‘investment value’ of €9.5bn and an ‘economic value’ (market size in money) of €68bn, whereas for
the United States these values amounted to €19bn and a staggering €750bn, respectively.
Accordingly, the main message of PIRA is that compared to the EU, the United States has only twice
the investment value for PSI but earns more than forty times from it. Although the PIRA figures were
disputed later, the report paved the way for the first PSI Directive, providing the European
(economic) rationale to legislate this matter.

MEPSIR
In 2006, seeking to benchmark the impact of the implementation of the PSI Directive, the European
Commission assigned the MEPSIR study19 to a Northern Ireland-Dutch consortium of Helm and Zenc.
This study undertook a thorough baseline measurement of PSI re-use across Europe (including
Norway), covering all major PSI sectors (but excluding scientific/research information and cultural
information). Although the MEPSIR study came up with much lower figures than the PIRA study – a
market size of €27 bn, it nevertheless confirmed the value potential inside PSI.20 The MEPSIR study
relied on a large number of robust measurements from all PSI domains in all Member States and it is
generally regarded as the best estimate.

MICUS
In 2008, a follow up benchmark study was carried out by the German consultancy firm MICUS in
three specific PSI domains, again under assignment from the EC. It reported that the PSI Directive has
had its strongest impact in the geographical information sector, followed by the legal and
administrative information market, whereas its impact in the meteorological information sector was
limited. Unfortunately, due to low response figures it did not do a second base measurement, which
would have allowed for a comparison with the MEPSIR figures.

Vickery
Then, in the autumn of 2011, upon request from the EC, Graham Vickery, former economist of the
OECD, produced a report, again assessing the overall European market, based on the various figures
presented in previous studies. Covering 27 EU Member States, the report assessed that (a) the
market size of 2008 and of 2010 amount to €28bn and €32bn, respectively, (b) the market features
an average growth rate of 7%, (c) total direct and indirect economic impact of PSI re-use lies between
€70bn and €140bn and (d) the welfare gains from moving to marginal cost pricing reach up to €40bn.

19
http://ec.europa.eu/information_society/policy/psi/docs/pdfs/mepsir/executive_summary.pdf
20
PIRA used a markedly different approach from MEPSIR. PIRA relied on two distinctively different values: an
‘investment value’ (public sector investments in the acquisition of PSI) of €9.5bn and an ‘economic value’ (part
of national income attributable to industries and activities built on the exploitation of PSI) of €68bn. Its
estimation of the investment value was based on a limited number of in-depth studies. Consequently, the
individual values of PIRA might be more robust but the subsequent aggregated value less robust. The
estimation of the economic value on the other hand was based on information derived from national accounts.
This implies a rather broad definition of market size: all firms which are in one way or another related to PSI.
The numbers in the MEPSIR study were solely based on the total added value by all first-order re-users (based
on a much larger number of measurements than PIRA), as it considered the heart of the matter how much of
the added value can be traced back to PSI (and not whether the information industry represents a significant
part of a national economy (as is the case in the USA)). This (likely) explains why the PIRA base value (€68bn) is
so much higher and the range so much wider (€28bn to €134bn).
22

The table below provides an overview of these studies and their main figures.

21
Year Name of study and PSI domain Outline
author
2000 Commercial Exploitation of - All PSI domains - For Europe (15 Member States)
Europe’s - 15 EU Member o Investment value: €9.5bn
Public Sector Information, States o Economic value (=market size):
22
Pira International €68bn
- For the United States:
o Investment value: €19bn
o Economic value: €750bn
2006 MEPSIR, - All PSI domains - European market size: €27bn (27 EU
Measuring European - 27 EU Member Member States + Norway)
Public Sector States + Norway
Information Resources,
HELM and ZENC
2008 Assessment of the - Geographic The PSI Directive has had strongest impact
Re-use of PSI in the information, in the geographical information sector,
Geographical Information, - Meteorological followed by the growing legal and
Meteorological information, administrative information market. Its
Information and - Legal information impact has been limited in the
Legal Information - 27 EU Member meteorological information sector.
23
Sectors, MICUS States
2011 Review of Recent - All PSI domains - Market size 2008 and 2010: €28bn
Studies on PSI Re-use - 27 EU Member and €32bn
and Related Market States - Average growth rate in PSI-related
Developments, Graham markets: 7%

24
Vickery - Total direct and indirect economic
impact of PSI re-use: €70 bn–€140bn
- Welfare gains from moving to
marginal cost pricing: €40bn
Figure 3-1: Overview of PSI case studies

In summary, all the studies acknowledged the economic value captured in PSI and the significant
growth rates over the years when opened up under liberal re-use regimes. In the next paragraph, we
will take a closer look at these effects.

3.3 Overview case studies

Case studies: evidence building up
As indicated above, the PSI Directive was adopted on 31 December 2003, putting the burden on
Member States to implement the provisions into national legislation by July 2005. In that process,
many Member States started to understand and appreciate the potential of opening up their PSI, in
particular the United Kingdom, where the so called Trading Funds were a sore subject to many re-
users (and likely also policy makers). Interest in PSI re-use was further boosted by the ‘Open Data
movement’, which started to gain political weight in 2010, bringing forth a second wave of studies of
which POPSIS, Koski and Houghton are of particular interest. The table below provides a (large)
selection of relevant studies, and briefly outlines the essence. For those interested in further details,

21
Meaning: year of publication.
22
http://ec.europa.eu/information_society/policy/psi/docs/pdfs/pira_study/commercial_final_report.pdf
23
http://www.epsiplatform.eu/media/files/micus_studie_psi_eu_long
24
http://epsiplatform.eu/content/review-recent-psi-re-use-studies-published
23

structured summaries of all these case studies have been brought together in Annex 4. Furthermore,
the bibliography in Annex 7 provides a full overview of relevant material.

Year Title and author + short PSI domain (+ short names) Outline
name
2008 ‘Models of Public Sector A set of UK ‘basic registers’: Relying on prior experiences of agencies
Information Provision via - Met Office adopting marginal cost pricing, the study
Trading Funds’, Newbery, - Ordnance Survey provides estimates for the costs and
Bentley and Pollock, - Hydrographic Office benefits of marginal cost pricing in
Cambridge University, - Land Registry relation to bulk, digital PSI from big UK
25
Trading Funds study - Companies House public data holders.
- Driver Vehicle Licensing
Agency
2009 ‘The Economics of Public UK raw PSI in general Relying on mathematical analysis the
Sector Information’, study assesses who should best finance
Rufus Pollock, Cambridge PSI re-use and the regulatory structure
University, needed.
26
Pollock study
2010 ‘PSI in European Meteorological information Proceeding on the basis that, in general,
Meteorology – an in general meteorological PSI is available on a cost-
Unfulfilled Potential’, recovery basis in Europe and on
Richard Pettifer, PRIMET, marginal or zero cost bases in the US,
27
Pettifer 1 study the study assesses the detrimental
effects for Europe.
2011 ‘Pricing of Public Sector 21 case studies in the EU in Analysing 21 case studies, covering a
Information Study’, all important PSI domains, wide range of PSBs and different PSI
Deloitte Belgium, including, the Dutch KNMI sectors, the study assesses different
28
POPSIS study case (meteo), Norwegian models of supply and charging for PSI

MET.NO case (meteo), and their effects on the downstream
Danish Deca case market, PSI re-users, end-users and
(geographic), Spanish impacts on the PSB itself.
cadastre case (geographic),
Austrian cadastre case
(geographic)
2011 ‘Does Marginal Cost Geographic information Assessing the performance of 14,000
Pricing of Public Sector firms in the architectural, engineering
Information Spur Firm and related technical consultancy
Growth?’ Heli Koski, The sectors, located in 15 different
Research Institute of the countries, the study analyses the effect
Finnish Economy, of maximum marginal cost pricing for
29
Koski study geographical PSI on the firms’ growth
performance during the years 2000–
2007.

25
http://www.berr.gov.uk/files/file45136.pdf
26
http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe0920.pdf
27
http://www.primet.org/file/EU%20PSI%20Working%20Groups/PSI%20in%20European%20Meteorology%20-
%20an%20unfulfilled%20potential%20distribution%20copy.pdf
28
http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012/models.pdf
29
http://www.etla.fi/files/2696_no_1260.pdf
24

Year Title and author + short PSI domain (+ short names) Outline
name
2011 Costs and Benefits of Information from: Presenting three case studies, the study
Data Provision – Report - Australian Bureau of explores the costs and benefits that PSI-
to the Australian Statistics producing agencies and their users
National Data Service, - Office of Spatial Data experience in making information freely
John Houghton, Victoria Management & available and the preliminary estimates
University, Australia, Geoscience of the wider economic impacts of open
30
Houghton study - National Water access to PSI.
Commission & Bureau of
Meteorology
2011 ‘Pricing of PSI in the Meteorological information Considering three hypothetical SMEs, in
Meteorological Sector in general Luxembourg, Poland and France that
blocks market wish to provide weather-related services
development’, Richard but are confronted with cost-recovery
Pettifer, PRIMET, pricing, the study concludes that these
31
Pettifer 2 study SMEs can never compete successfully.
Figure 3-2: Overview of relevant PSI re-use case studies

In summary, we see a surge in interest in the economic effects brought about by more liberal PSI re-
use regimes, mirrored by a wide array of studies at the national and sectoral levels, both inside and
outside Europe. In the next paragraph, we will look at these effects in more detail and will refer to
these case studies by using their short names.

3.4 Chain of economic effects of lowered PSI re-use charges

The challenge: putting the findings in context
Most of the studies referred to above work on the basis of an input–output relation, whereby the

effects (output) of lowered charges for PSI (input) are assessed in isolation. What has not been done
so far is to analyse the studies by interconnecting them and adding a sequence and timeframe to the
effects, which will not only reveal the deadlock faced but also suggest the solutions at hand.

The sequence of effects and their beneficiaries
Our starting point (and that of most case studies) is a PSB lowering the charges for re-use of its PSI.
This brings about a whole array of subsequent effects, which can be divided in three phases: (a) the
sowing phase, (b) the growing phase and (c) the harvesting phase. Walking through, we will look at
the subsequent effects taking place and substantiate them with the empirical research listed above.

a) The sowing phase
The sowing phase features two immediate effects: an uptake at the demand side by the re-users
(following the price cut) and, in parallel, an income effect at the side of the PSB, where it loses
revenues – it can no longer charge for its data – and sees its cost increasing, as mirrored in the
illustration below.

30
http://ands.org.au/resource/houghton-cost-benefit-study.pdf
http://www.crcsi.com.au/Documents/ANZLIC-Economic-Study---Stage-2-Report.aspx
31
http://www.primet.org/file/EU PSI Working Groups/Workshop Position paper final
25

Sowing Phase 2 Phase 3
Benefits
Growing Harvesting
created in phase
euros
• Business • Welfare effects
• Demand effect re-
effect re- users
users Non-economic
• Market
benefits
dynamism
effect
Benefits
private sector

Time
0
X1 X2 ∞
Costs/missed
revenues
Costs • Income public sector
incurred • Efficiency • Taxation effect
effect PSB effect PSB
in euros

Figure 3-3: Effects in the sowing phase

i. Demand effect re-users

Spectacular increases in demand
All case studies report on, quite often really spectacular, increases in demand (both in terms of
volume and numbers of users) following a decrease in re-use charges, as demonstrated in the table
below.

Interestingly, lowered prices also attract new categories of users, SMEs in particular, apparently
previously unable to afford the required PSI. Both POPSIS and Koski report on this:

POPSIS: “Interestingly, some case studies demonstrate the use of variable pricing regimes such as
‘pay per use’ or ‘percentage of turnover generated by PSI’ without high fixed price elements. These
regimes have led to increased re-use and facilitate new entrance of re-users, notably SMEs.”32

Koski: “It seems credible that higher PSI prices create a barrier for SMEs using geographical
information to develop new information products and services and to enter new market areas.”33

We will look at the consequences thereof in more detail in the growing phase, where they
materialize in full.

32
POPSIS, ibid. p. 32.
33
Koski, ibid. p. 13.
26

Case study PSI domain Price cut re- Increase in demand
use charges
Austrian Cadastre Topographical data Up to 97% Factor 2 – 7 in number of downloads34
(POPSIS + Koski)
DECA (POPSIS) Danish address data almost 100% Factor 100 in number of re-users
KNMI (POPSIS) Dutch meteo data 80% Factor 10 in number of re-users, 90%
of them being SMEs
MET.NO (POPSIS) Norwegian meteo 100% Factor 30 in numbers of unique
data weekly re-users, majority being SMEs
Spanish Cadastre Spanish 100% - Factor 80 – 100 in numbers of
(POPSIS + Koski) topographical data downloads
- Factor 25 in numbers of re-users
Houghton study Australian:
- Topographical - almost 100% - 172%
data - 100%
- Statistical data - Factor 3 in product downloads
- Hydrological data - 100%
- Factor 100 in data requests
- Factor 2 for extractions of re-use
Figure 3-4: Overview of increases in demand following lowered PSI re-use charges

PSI features relatively elastic demand
These figures confirm previous research (Trading Funds Study and Pollock Study) that suggested a
price elasticity of demand (PED) well above 1 (in absolute terms), meaning that in case a PSB lowers
its prices (so not dropping the charging all together) the relative increase in quantity outweighs the
relative discount, generating higher revenues than before.35

Pollock notes that evidence on price elasticity is limited, and its value will be determined by the
nature of the product at issue. Nevertheless, he estimates that elasticity is generally greater than 1,
and the range for the kinds of products that are the subject of this study is between 0.5 and 2.5.
According to the Trading Funds Study, elasticity of demand varies depending on the PSI, but for the
products associated with the PSI, average elasticity is estimated at between 1 and 2.

The POPSIS findings confirm this in the Austrian cadastre case and the Dutch KNMI case.

POPSIS: “[T]he Austrian Federal Office of Metrology and Surveying adopted a simplified and more
market-oriented PSI pricing approach with drastic price cuts of up to 97% within strict budget
constraints (there was no additional governmental funding). Due to the additional demand – notably
from SMEs – triggered by lower prices, PSI sales revenues and the associated cost-recovery ratio could
be kept stable or slightly increased. Without additional governmental funding, BEV could improve the
situation for re-use business and secure a wider use of its public data.”36

POPSIS: “In 1999, at the peak of competition between the commercial activities of the KNMI and the
private sector re-use activities, there were in essence two re-users of KNMI data. … About ten years
later, after full implementation of the new re-use policy, this picture had changed quite dramatically.
34
Cartographic products with a factor 2 – 15, digital ortho-images with a factor 70, digital cadastral map and
elevation model with factor 2.5, the digital landscape model with a factor 10.
35
The formula for the coefficient of price elasticity of demand (PED) is (dQ/Q)/(dP/P), whereby Q is the
quantity, P is the price and d is the changes therein. Generally, if PED for a good is relatively elastic (-∞ < Ed < -
1), the percentage change in quantity demanded is greater than that in price. Hence, when the price is
lowered, the total revenue increases.
36
POPSIS, ibid., p. 127.
27

In 2010, the price level of a full KNMI dataset went down by 80%, from 0.1 M EUR to 0.02 M EUR
(which included both license and distribution costs) and covered the facilitation of re-use costs only.
At the same time, the number of re-users exploded, increasing to 5037.”38

ii. Income effect PSB
Lowering of the charges directly impacts (negatively) the income of the PSB, as revenues no longer
come in. In fact, as the new charging regime needs to be implemented, costs may further rise, in
particular as boosted demand may require additional investment. This double-edged knife requires
the PSB to rely on its own reserves and, in the absence thereof, requires alternative funding (from
general taxation funds).

Lost revenues appear to be limited in size
Interestingly, however, looking at the lost revenues in proportion to the PSB’s total budget, in most
cases the ‘damages’ appear to be fairly limited. The table below – directly taken from the POPSIS
39
study – indicates that the ‘PSI re-use cost-recovery ratio’ of more than half of the PSBs is less than
5%. In other words, if charges were dropped all together many would hardly notice, or at least, they
would not have to shut shop. Only for a few the loss of income would appear to be of a fundamental
nature, in particular in the field of business registers.
Country Public sector body PSI domain Budget (M PSI sales Cost-
EUR) revenues recovery
(M EUR) ratio
Business 93.6 31
Italy Infocamere register 31.31%
Business 243 6
Netherlands KvK register 19.50%
Business 74.8 15.5
United Kingdom Companies House register 20.73%

Geographic 85.0 22.5
Austria BEV information 26.5%
Germany BKG information 0.24%
Germany SenStadt information 10.38%
Denmark DECA information 0.82%
Geographic 52.0 2.1
Spain IGN-CENIG information 4.12%
Geographic 108.0 0
Spain Spanish Cadastre information 0.00%
France French cadastre information 0.55%
Italy Italian cadastre information 0.50%
Netherlands Dutch cadastre information 6.57%

37
Among these fifty companies, five are companies that are so-called meteorological service providers (in 1999
there were just two). These companies have portfolios with direct meteorological forecasting products as basic
meteorological datasets for customer processes (sea forecasting, wind energy and so on). General re-users deal
with customer processes that are built on meteorological input datasets.
38
POPSIS, ibid. pp. 273–274
39
The cost-recovery ratio is defined as: (PSB’s revenues from sale of raw PSI to re-users / total budget of the
PSB) * 100%
28

Country Public sector body PSI domain Budget (M PSI sales Cost-
EUR) revenues recovery
(M EUR) ratio
Geographic 127.0 21
United Kingdom Ordnance Survey information 16.54%
Meteorological 214.9 2
Germany DWD information 0.93%
Meteorological 56.0 0.25
Netherlands KNMI information 0.45%
Meteorological 58.0 0
Norway Met.no information 0.00%
Meteorological 6.0 0.36
Slovenia ARSO information 6.00%
Spain CENDOJ Legal 9.0 1.5
information 16.67%
France DILA Legal 135.0 0.9
information 0.67%
France SIRCOM Fuel prices 1.1 0.179
information 15.91%
Germany DeStatis Statistical 177.7 0.2
information 0.11%
Figure 3-5: Cost-recovery ratios 2010 of PSBs measured in the POPSIS study40

Equally, Houghton concludes that the greatest cost to agencies lies in the loss of revenue when
information that was previously sold is provided at marginal cost. Moving to standard licences and
formats may have some transitioning costs for agencies but is unlikely to have a material impact on

costs once the standard systems are in place. Ultimately, the use of standard licences and formats
should reduce agency costs by reducing the support required by re-users.

b) The growing phase
In the growing phase, more indirect effects start to kick in, both at the market end (for the re-users
and more broadly on the downstream market) as well as for the PSB itself, in the form of increased
efficiency, as the figure below illustrates.

40
Although only six cases from the POPSIS study were explicitly mentioned above, the table demonstrates all
21 POPSIS’s case studies, as the percentages are quite illustrative.
29

Sowing Growing Phase 3
Benefits
Harvesting
created in phase phase
euros
• Welfare effects
• Demand • Business Non-economic
effect re- effect re- benefits
users users
Benefits
• Market private sector
dynamism
effect Additional tax
incomes

0 Time
X1 X2 ∞
Costs/missed
revenues
public sector
Costs • Income • Efficiency
incurred • Taxation effect
effect PSB effect PSB
in euros

Figure 3-6: Effects in the growing phase

iii. Business effect for re-users
The costs of purchasing PSI from the government will decrease, which is (partly) translated into
lowered prices in the successive parts of the chain, leading to larger quantities sold, and as the price
cuts do not affect profits (but are a result of lowered costs), profits of re-users rise. Furthermore,
transaction costs diminish, in case re-use is made free all together.

Koski has looked extensively at profitability of companies following lowered re-use charges.
Assessing the performance of 14,000 firms in the architectural, engineering and related technical
consultancy sectors, located in 15 countries, she analyses the effect of maximum marginal cost
pricing for geographical PSI on the firms’ growth performance during the years 2000–2007.

Koski: “The reported empirical findings clearly show that the PSI pricing scheme does matter for the
firm growth particularly from the perspective of small and medium sized enterprises. The firm-level
data concerning potential re-users of geographical information in business services sector from 15
countries during the years 2000–2007 suggests that the pricing of GI strongly relates to the firms’
sales growth. Firms functioning in the countries in which public sector agencies provide fundamental
geographical information either freely or at maximum marginal costs have grown, on average, 15
percent more per annum than the firms in the countries in which public sector GI is priced according
to the cost-recovery principles. The difference-in-difference estimations further show that positive
growth impact materializes already one year after switching to the marginal cost pricing scheme but
a stronger boost to the firm growth takes place with a two year lag.”41

41
Koski, ibid. p.13.
30

Three POPSIS cases also report increased turnover of re-users following a move to marginal costs
charging models: in the Danish address case (DECA), the turnover of re-use market (first and second
tier re-users) increased by 1,000%; in the Dutch meteo case (KNMI), the turnover of the downstream
market increased by 400%; and in the Norwegian re-use meteo market case (MET.NO) there was a
200% growth, money-wise. Many case studies also demonstrate new parties entering the market,
which is obviously the result of increased profitability (see Figure 3-6 above). We will look at these
new entrants in more detail in the next section (market dynamism).

iv. Market dynamism effect
Attracted by low (or non-existent) PSI re-use charges, lowering market entry barriers, and increasing
profits of existing re-users, new parties enter the market, resulting in more market dynamism:
existing re-users need to innovate and upgrade their services. Conversely, parties no longer adding
value and not able to keep pace, leave the market.

SMEs entering the market spur dynamism
Koski’s research demonstrates that the market dynamism and growth is spurred by the new comers
rather than the existing body of re-users, bringing about a subsequent set of economic effects. This is
confirmed by a wide range of case studies.

Koski: “Interestingly, marginal cost pricing has not generated notable growth among the large firms;
it has been SMEs that have benefited most from cheaper geographical information. It seems credible
that higher PSI prices create a barrier for SMEs using geographical information to develop new
information products and services and to enter new market areas. The switch to the marginal cost
pricing may thus not only result in growing markets but also intensify competition and challenge the
large incumbent companies. Cheaper public sector GI is thus likely to benefit consumers by producing
more product variety and also cheaper prices.”42

Distinction between high- and low-end markets
POPSIS addresses another important aspect of the market dynamism brought about. It notes that
‘the new kids on the block knocking on the door’ are fundamentally different from the ‘old’ re-users
(the high-end market re-users), as Figure 3-7 demonstrates.43

Figure 3-7: High-end and low-end re-use markets

The high-end market typically consists of re-users that provide their PSI-based services to
professional clients. Substantial value is added by re-users serving the needs of specific professional
clients. A typical example is a meteorological company that provides very detailed weather forecasts

42
Koski, ibid. p.13.
43
POPSIS, ibid. pp. 26–27.
31

to oil rigs, based on its own high-tech forecast models. The high-end market services are highly
targeted, the number of clients is relatively low and yet the value of each transaction is high.

Conversely, in the low-end market, business models are based on reaching out to large volumes of
(generally non-professional) consumers who use high traffic web services and maybe apps on mobile
devices. Typically, these re-users merely mash up the PSI with other free content and integrate it into
services, not adding much value, other than distributing it widely. The re-users’ revenues come from
third-party advertisements, not from its users.

Backing this up, in the Dutch KNMI case, the POPSIS study reports:
“The lowered price level increased competition and sparked innovation: second-tier users of
meteorological information were offered smart, new products. For instance, the greenhouse sector in
the Netherlands was able to save about 10% on its energy costs due to its access to real-time detailed
forecasting of rainfall services. This allowed the sector to maximize the length of time that the
greenhouses can remain open to the air. This not only very beneficial for the crops but also for the
environment, since it reduces carbon dioxide emission quite considerably.... New business models
emerged: a new re-user entered the market and launched an innovative service under the name
‘Rainfall Radar’ (Buienradar). Anyone can use the service to determine whether it is going to rain in
the current location in the next few hours. This service is provided completely free of charge. It
generated around 300 million hits per year throughout Europe in 2010. As a result of this high traffic,
it is paid for through advertising revenues. Finally, since all KNMI data products are license free,
almost no restrictions in use or distribution are set. Some of the re-users have started activities as
distributors.”44

Stifling effects of strong PSB presence in the market
One may also ask what the consequences are when prices are not lowered. The two Pettifer studies
provide evidence of the consequent damage being incurred in the meteo domain: not lowering its re-

use charges and its own downstream market activities, the national Met Offices trifles with the
market, where high charges block SMEs from entering.

In his first paper ‘Pricing of PSI in the Meteorological Sector blocks market development’, Pettifer
considers three hypothetical SMEs, in Luxembourg, Poland and France that provide weather-related
services relating to forecasting, highways and energy, and uses 2010 prices on a cost-recovery basis.
The absolute minimum PSI meteorological data required to provide basic weather-related services,
with a market value of €6 ,000 to €20,000 per contract, would cost a typical SME between €84,000
and €400,000. Pettifer’s conclusion therefore is that SMEs cannot operate successfully or compete
with large firms when partial or full cost-recovery pricing principles are used: cost-recovery pricing
principles are likely to create barriers to market entry because SMEs are probably unable to find the
20 contracts required to operate profitably.

In Pettifer’s second paper, ‘PSI in European Meteorology – an Unfulfilled Potential’, he assesses the
damage from this current practice, by comparing the European market figures with those of the US.
“Recent estimates of the size of the 2006 market in value-added meteorological products of all
types in the USA and Europe are of the order of $1.4 billion per annum and $372 million (€530 million)
per annum respectively. ... It would appear therefore that [on the basis of GDP] only about 0.3% of
the potential European market in this sector is currently being supplied whereas in the US the
equivalent figure is around 0.7%. Moreover recent estimates suggest that in real terms, after allowing
for growth in GDP, the US market has grown at an average rate of around 17% per annum over the
past six or seven years while the European market has been growing at closer to 1.2% per annum in
the same period. This type of difference can be seen in specific market sectors as well as in the overall

44
POPSIS, ibid. p. 274.
32

market. ... There are other characteristics of the European meteorological market that bear
examination and raise questions over the structure and operation of the sector. For example,
although the real overall annual market growth in Europe has been languishing below 2% over the
past five years, the small part of it (now about 28%) that falls to the private sector has been growing
at around 25% per annum whereas the 75% that is in the hands of the dominant NMHS has actually
declined by around 1.5% per annum. This large growth in the private sector component of the market
(albeit from a very low base) is doubtless to some extent due to capture of business from the NMHS
but the NMHS, despite their greater resources and strong brand positions, appear unable to develop
the market and to grow the meteorological economy overall. It is interesting to note that much of this
increase in the private sector component of the market has arisen since a few of the NHMS relaxed
their PSI supply policies partly or completely towards the US model and made some key
meteorological PSI available at the marginal cost of distribution. This suggests that if a major overall
structural change in this direction, whether political or commercial, can be made it will encourage the
growth of the private sector, stimulate genuine competition and foster the development of the total
market.”45

v. Efficiency effects in the PSB
The last effect in the growing phase concerns the efficiency gains made by the PSB. Having
implemented the new re-use policy, efficiency gains start to kick in where fewer resources are
consumed to run operations, administrative staff are no longer needed and transaction costs savings
are cashed in. Furthermore, the PSB’s remaining downstream market activities are seized or carved
out, as the PSB cannot keep up with private sector competitors (no longer being able to rely on
upstream advantages).

Direct efficiency gains cover significant part of revenues lost
Houghton compares the PSB’s revenue lost with the costs saved as a result thereof: (Efficiency gains
PSB / Lost income PSB) * 100%. Accordingly he arrives at a percentage of 32% (for the Australian

Bureau of Statistics). He also reports on Western Australia’s Landgate agency, which estimates their
transaction and support-related costs at around 17% of fundamental data revenue.46

POPSIS reports on several cases where efficiency gains are being perceived:

 The Dutch KNMI case demonstrates the move to a marginal costing charging regime had a
significant impact on the efficiency of the organisation. In 1999, the commercial arm of the
KNMI comprised 25 FTEs. This amounted to a cost of around €0.65m (in both direct and
indirect costs) and a turnover of PSI sales of about the same amount, so breaking even. In
2010, the ‘re-use department’ of the KNMI ran at a total cost of around €0.25m a year by 1.5
FTE, whereby the number of re-users had gone up with a factor 10.

 The Danish address case (DECA) assessed that opening up address data against marginal cost
led to a total savings of around €5m over a period of five years, by estimating the time saved
from not having to deal with licensing and administrative issues anymore. That amount alone
already outweighed the loss of PSB income. The re-use department is now run by 0.5 FTE
against a cost of €0.2m.

c) The harvesting phase
As Figure 3-8 demonstrates, in the harvesting phase the investments really pay off: increased tax
returns outweigh the costs incurred by the PSBs and market dynamism has led to economic growth,

45
Pettifer, ibid. pp. 5–6.
46
Houghton, Ibid. pp. 19, 34.
33

resulting in more employment. And of course there are other non-economic benefits, all adding up
to increased welfare.

Benefits Sowing Growing Harvesting
Benefits
created in phase phase phase private sector
euros
Non-economic
• Demand • Business • Employment effect benefits
effect re- effect re- • Welfare effect
users users Additional tax
• Market incomes
dynamism
Income
effect Treasury

0 Time
X1 X2 ∞
Costs/missed
revenues
public sector
Costs • Income effect • Efficiency • Taxation effect
incurred PSB effect PSB
in euros

Figure 3-8: Effects in the harvesting phase

vi. Employment effect
Due to increased economic activities, new and old re-users (and their clients) require additional staff,
which they can now afford to hire due to increased profitability. This positive effect on employment
largely outweighs the possible redundancy at the PSB level.

Although the strong increase in usage and the numbers of re-users suggest that there must be
positive effects on private sector employment, only two cases hold hard evidence. The Dutch KNMI
case demonstrates that in the eleven years since 1999 following the policy change that entailed the
shift to cost recovery of re-use facilitation costs only, there were significant developments. The
number of re-users went up by 1,000%, turnover increased by 400% and employment was boosted
by 300%. In the Danish DECA address data case, which also shifted to a re-use facilitation cost-
recovery model, the number of re-users went up by at least 5,400%, turnover by 1,000% and
employment by first and second tier re-users by 800%.

vii. Taxation effect
Although already taking off in phase 2, the treasury starts to benefit from the entire movement: (a)
VAT returns increase, (b) profit taxation increases as the GDP goes up and (c) social security taxes
increase where more people are employed. These returns start to outweigh the initial investments
made, having to fund the budget gap of the PSB in phase 1 and part of phase 2.

Again, few cases hold hard figures on the taxation effects. In the KNMI case, the additional corporate
tax gains amount to €35m over a period of eleven years, based on an ‘investment’ (= lost revenues +

34

re-use facilitation costs) of around €7m , thus giving a return on investment of 500%. The Danish
DECA address case suggests a similar return on investment (corporate tax only) of 450%, where the
GDP increase amounts to €14.25m over nine years against an investment of €3m. Obviously, at a
macro level, and in absolute terms, these amounts are modest but become significant when scaled
to a European level.

Pettifer, in his paper, ‘PSI in European Meteorology – an Unfulfilled Potential, points to the potential
tax gains missed out on:
“The failure to realize the potential in this market place is costing the national treasuries in the EU
dearly in terms of lost revenue from taxation. If the European meteorological market were as well
penetrated as that of the USA, then the actual market size would be around €1,390M per annum.
According to Eurostat the overall taxation return for EU countries in 2005 was 39.6% of GDP. To a
first approximation then we might expect that the gross overall tax revenue from this sector would
increase by around €340M. If, to generate this, the NMHS were to lose all of their income from the
sale of PSI, and all of their direct value added retail sales (which are assumed to be diverted to the
private sector and are thus still within the total market size), then the net benefit to the EU central
treasuries from this change in the trading structure of the market would be in the order of €290Mper
annum and would be, if the US is any guide, growing at about 17% per annum in real terms, rather
than at about 1.2% per annum as they now are.”

Treasury implications
Returning to Figure 3-8, the yellow and green planes represent the cash implications for the
Treasury. During the sowing phase, the loss of income and investments needed to kick start and
implement free re-use are not set off yet by taxation gains, which start to kick in only in the growing
phase. At point x1 the positive cash flow outweighs the loss of income and accordingly, the cash
needs accumulated in the previous period (0 – x1) start to diminish rapidly. At point x2 the net gains
start to come in, where these cash needs accrued are outweighed by the positive effects (tax gains

and PSB efficiency gains). As of this point, the investment decision towards a free or marginal costs
charging model will yield a constant and structural return.

0

So the balance of Treasury = X2 f(Additional tax incomes) – f(Costs + missed incomes public sector)
where:
x1 x
= point where tax revenues start to outweigh the PSB’s total costs and income lost
x2 = point where total tax revenues collected outweigh the PSB’s total costs and income lost
x2 - ∞ = total taxation profits

So ultimately, the cases looked into promise high returns. However, where PSBs have become reliant
on income from re-use, the turnaround may not be easy, particularly if the proportion of user fees is
relatively high. In those cases the Treasury will need to finance the transition, making up for the
initial losses. We will look into this matter in more detail in chapters 4 and 5.

viii. Welfare effects
Finally, it is likely that the societal gains will be much higher than just the financial ones we have
been focusing on. 47 Many reports underline the positive externalities from opening up PSI (for re-
use), which will likely result in wider economic impacts and benefits for society or the public at large.
The benefits may look trivial – e.g. localized weather forecasts help people to stay dry – but they are

47
See in particular Velde te, R.A. (2009), Public Sector Information: Why bother? in: Uhlir, P. (ed.). The socio-
economic effects of Public Sector Information on Digital Networks. Towards a better understanding of different
access and reuse policies. Washington DC: National Research Council (Ch.6. pp.25–28) and on the same note:
http://www.oecd.org/dataoecd/12/49/40064800.pdf
35

obvious: less flu, more productivity, happier citizens, etc. Equally, we have only started to understand
the potential network effects of opening up PSI, in terms of innovation and the development and
introduction of new products, services and processes that, in turn, generate new economic activity,
new business opportunities, better informed and potentially better government and business
decisions. Making sure we do not forget, these effects are represented by the blue dotted line in
Figure 3-8.

3.5 Wrap up

Empirical research in various PSI domains seems to confirm the theoretical assumptions reflected in
chapter 2. Briefly put, the benefits largely outweigh the costs (and lost incomes of the PSBs): public
sector activities increase, leading to economic growth, more employment and better services, due to
market dynamism. The additional tax returns make up for the extra costs and income loss of the
PSBs. The only issue is that these macro returns only kick in later, requiring transitory financing
arrangements.


36

4 WOULD IT WORK WITH US? – TRANSPOSITION IN THE GMES
DOMAIN

4.1 Introduction
In the previous chapter we saw that there is a growing base of evidence that from a macroeconomic
perspective, a free data re-use policy makes good sense. In paragraph 4.2 we will look at evidence
coming from within the GMES domain (or more practically, the wider EO domain), allowing us to
compare it with the empirical evidence gathered from within the other PSI domains and letting us
make an estimate of the potential benefits of a free GMES data re-use policy in paragraph 4.3, which
takes the form of an educated guess since the quantitative data are currently limited.

4.2 EO domain evidence
Although fairly limited there are a few cases that allow us to look at the economic effects of changes
in EO data re-use policies, being: (a) Landsat, (b) ERS and Envisat, (c) examples of national EO data
portals opening up for free re-use and (d) some anecdotal facts and figures from downstream
companies relying on EO data.

a. Landsat – the long transition from commercial to free
The clearest example is the US satellite system, Landsat. It is the best case because during its 40
years of imaging, using a succession of satellites, it has been subject to several changes of data
policy. Most recently, in 2008 it moved from a charging to a free-access data policy and the impacts
of this last change are clearly visible from the increase in data downloads as shown in Figure 4-1.

It is also notable that the change comes after years of charges and many changes in policy. Since

commercialisation was first tested in 1981, the operational responsibility for the satellites has
changed hands three times and many different charging structures have been tried out. In the mid-
1980s a new company was established to sell the data at a price that peaked at $4400. This policy
was deemed to have failed when, in the early 1990s, it was recognised that:

“With military and government-funded re-searchers as major Landsat data users, Congress—and the
public, via a spate of news-paper articles—realized it was “paying twice” for Landsat data, in that tax
dollars had built the satellites and then by paying a government-subsidized company for its data. As a
result, after many Congressional hearings the decade-long commercialization experiment was ended,
and the Landsat program returned to the government fold.”48

This led to a decade where the data was sold directly by the US Geological Survey (USGS) at a price of
$600 per scene until, in 2008, the US Government decided that data gathered by the Landsat system
of satellites should be made available at no charge to users. Almost overnight the number of
downloads of imagery exploded. 49 In a recent analysis of the impact of the data policy it was noted
that:

“Open access has resulted in the distribution of over 5.7 million images (through June 2011),
representing the full range of Landsat instruments. More than 250,000 images are distributed each

48
Chronicling the Landsat Legacy Laura Rocchio, SSAI; The Earth Observer, Nov-Dec 2011
49
Landsat Data Distribution Policy, January 2008,
https://landsat.usgs.gov/documents/Landsat_Data_Policy.pdf
37

month—an incredible statistic when considering that for the entire year of 2001 (when the previous
record was set for data distribution) approximately 25,000 images were purchased.”50

In 2001, the year of highest downloads before 2008, nearly 20,000 scenes were taken at an average
of 53 scenes per day. In 2011, over 2 million scenes were downloaded at an average of over 5,000
per day; a 100-fold increase. Clearly the interest in the data is very strong. See Figure 4-1 below.

Figure 4-1: Landsat image downloads since 2008 (courtesy of USGS)

For a full history of the Landsat programme see [Wulder et al50 and USGS54].

Economic effects
In paragraph 3.4 we saw a series of economic effects resulting from the opening up of PSI in the
other domains. How do these compare to the Landsat case?

i) Demand effect
This dramatic increase in the use of the data corresponds with the findings in other domains seen in
the sowing phase (see paragraph 3.4). As the charging policy is changed and data is made freely
available, the interest increases by one or two orders of magnitude. There appears to be a definite
read-across from other PSI domains to the EO domain.

ii) Efficiency effect
The efficiency effect is where the PSB no longer has the cost overhead of maintaining the resources
needed to run the sales operation. In this respect, the USGS 2011 report says that “because the
internet makes it possible for users to download images directly … the bureau has realised savings
including eliminating the billing and accounting system.”51

50
Wulder, M.A., et al., Opening the archive: How free data has enabled the science and monitoring promise of
Landsat, Remote Sensing of Environment (2012), doi:10.1016/j.rse.2012.01.010
51
Landsat Fees, from the USGS website http://remotesensing.usgs.gov/landsat_fees.php
38

As in the other domains, the income realised by the USGS was very low and barely covered the cost
of the sales system.

iii) Market dynamism effect
The users of Landsat imagery especially in the private sector have always been very sensitive to the
price demanded. Recall that in 1985, a single uncorrected Landsat scene cost $4,400, in 2001 it
would have been $600 and today it is free. The 1986 annual report on Landsat sales reported that:

“[T]he announcements in 1981, 1983 and 1985 that the prices would rise led to a roughly doubling of
demand in the months preceding the increase and to a drop in sales to very low sales immediately
following the increase52. [Furthermore,] in 1983 when prices tripled, government purchase revenues
increased 5 fold whilst the value of private purchases fell by one third”53.

Nevertheless, when the USGS made a survey of the users of the data in 2011, it found that “on
average, respondents were willing to pay $760 per scene which is greater than the previously
administratively set price.”54

This apparently rather confusing and somewhat surprising result provides evidence of the sensitivity
of entry barriers for commercial users. Since the majority of those downloading Landsat data are
“public good users”, it follows that the largest drop is from the commercial sector. Charging creates
too high an entry barrier and it is probably reasonable to conclude that the strongest impact falls on
SMEs. Certainly, intuitively, this would be the case.

Furthermore, the USGS study54 shows that it is the academic and federal government users that are
willing to pay the highest prices. As these are the largest users, in the set-up where Landsat charges
were made, most of the funds were simply circulating within government circles.

Cutting the charges spurs innovation and re-use and creates new markets for the information.

iv) Business effect.
The strong surges in data downloads provide evidence that the data has value. Are there cases where
users of the data can show improved business results? In the course of the study we spoke with two
of the larger commercial users of Landsat data, the essence thereof is summarized below.

 MDA Federal is a private company and one of the largest users of Landsat data. They report
that out of $30m annual revenues, over $10m are generated using Landsat data and that as
a result of the policy change, they have recruited an additional 125 employees.
Furthermore, whilst they used to have a business in correcting the data, this has been
replaced by additional processing in the PSB. In other words, the PSB has moved forward in
the value chain in a way that the industry approves of! It is also a great advantage for
Landsat data that it is easy for users to download, partly helped by the enhanced degree of
processing ensuring geo-location accuracy. This is an important point to return to because
it is insufficient to just make data available free, it must also be easy to access. If it is not, all
savings in data purchase go in to the time and effort needed to access the data.

 GDA Corporation is another company to report on the benefits derived from Landsat’s
revised data policy. GDA Corp is a relatively young company, formed in 2002; it has about
10 employees. GDA offers a crop mapping service that is mainly used by federal and state
52
Encouraging Private Investment in Space Activities; Report to the US Congress, 1991.
53
USGS EROS data centre Annual report for Landsat sales 1987.
54
The users, uses and value of Landsat and Other Moderate-Resolution Satellite Imagery in the United States;
Miller et al. USGS 2011.
39

governments, particularly the US Department of Agriculture, Foreign Agriculture Service.
Whilst unwilling to share revenue figures, GDA considers that the value of no-cost data is
largely passed on to the customer through a cost saving of between $1.5m and $2m per
annum.

According to USGS: “Access to readily available Landsat imagery has helped GDA to expand rapidly
and serve agricultural, environmental, and resource management clients, some of whom were
previously unaware of the tremendous value that Landsat data and its analysis can bring to their
decision-making and ultimately their bottom line.”55

Hence, it would seem that the recent change in Landsat data re-use policy has acted as a strong
enabler driving business that probably would not have existed without a free data policy. A recent
paper56 from the Landsat advisory group underlines the potential for public sector benefits. Ten
application areas have been examined to give estimates of annual efficiency savings and are
considered to produce savings of at least $178m to $235m per year for the Federal and State
governments.

b. EO data re-use in Europe
Meanwhile, back in Europe, no situation compares with that for the US Landsat. Nevertheless, we
take the opportunity to include relevant figures since these do show a picture of what was happening
and have not been reported elsewhere. The closest comparison is with the SAR and MERIS sensors
carried on ERS-1, ERS-2 and Envisat. These three satellites, launched and operated by ESA between
1990s and 2012, were subject to a policy of commercialisation from the outset. A regime was
established that was similar to the US approach in the 1980s with one, and later two, companies
awarded the rights to sell the data.

In 2000, before the launch of Envisat in 2002, 2 commercial “Distributing Entities” were established

through competition. EMMA and SARCOM became responsible for the sales of ERS and Envisat data.
The total for sales - number of scenes - is shown for each year in the Figure 4-2 below.

14000

12000

10000

8000
Total Envisat
6000
Total ERS
4000

2000

0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011

Figure 4-2: Total sales (number of scenes) of ERS and Envisat SAR data.

55
Landsat adds tremendous value to decision-making and bottom line, USGS success story,
https://landsat.usgs.gov/Landsat_Stories.php
56
The Value Proposition for Ten Landsat Applications; Landsat Advisory Group, September 2012.
40

The peak in 2002 and 2003 is due to two exceptional bulk orders for ERS data. Further bulk orders
pushed sales up in 2007 and 2008 leaving the core sales more or less flat until ERS-2 was
decommissioned in mid-2011. Throughout the period, we see a steady rise in the sales of Envisat
data. Envisat failed suddenly in April 2012.

Regarding pricing, in 2000 the distributors paid 150Euro per scene (for bulk orders) to ESA and this
price was progressively reduced down to 12Euro in 2009. This pricing reflected both the reduced
price for Landsat data as well as the reduced cost of distribution. Revenues to ESA peaked at around
500k in 2002/3 and fell to under 100k in 2011. In this case, there seems to be no clear trend as has
been seen with Landsat.

However, in addition to commercial sales through the distributing entities, Envisat data was also
distributed at a marginal cost to research users. As costs fell and the new Landsat data policy
reduced prices, the cost of access to Envisat data was similarly reduced. This spurred interest and
seems to have led to a consequent increase in demand as shown in Figure 4-3. Note that the figures
are for projects using Envisat data and are not the number of scenes used. Hence these figures are
not comparable with those for Landsat data.


Figure 4-3: Elimination of barriers led to increase in data take-up for Envisat (courtesy of ESA)

Whilst less dramatic than is the case for Landsat, the uptake of Envisat data for user projects has
increased by a factor of 7 to 8 driven by easier access – although it is not possible to separate the
effects of technology change and data policy both of which played a role. Commercial sales have
changed much less and we may well be seeing some of these moving to project based work where
data costs were lower. Furthermore, there is no singular event, as in the case of the Landsat price
policy change in 2008 that triggered a change in demand.

c. National data portals
Further evidence is also building up at national level. At least two European Member States have
recognised the benefits that can come from the re-use of EO data. We describe in Annex 5, the cases

41

of Sweden and the Netherlands that have both created data portals containing sets of satellite data.
In each case, satellite data was being purchased from private operators by one government
department (but used by many more) before the decision was taken to make this same data freely
available to any user, including other government departments.

Both portals contain image sets of national coverage and hence interest. Both are aimed at national
users with the intention to stimulate local industry as well as public sector interest. Both require
users to register a national address but in reality the checks are limited. Moreover, since both
datasets have only national coverage, there is low interest for users coming from outside the
country; this is not the case for GMES data, which will have global coverage.

Since the data was being purchased, the licencing conditions were re-negotiated with the suppliers,
which involved some additional cost. Costs are covered by a consortium of public sector
organisations, but in both cases with the participation of some industries. In the case of Sweden, the
use of the data has increased dramatically by several 100-fold, where, over the longer time period,
2500 user accounts were reached in 2011.

In the Netherlands, where the policy change has been implemented just recently (March 2012), the
interest is very clear with an average of five new accounts being created each week. This now adds
up to around 250 accounts (with many more users behind that).

d. Downstream business
The value-adding (VA) sector in Europe has been estimated to have revenues of around €300m and a
cumulative annual growth rate (CAGR) of 7% according to the last industry survey that was made in
200657. New figures for the European market will be available after this report has been completed,
but Euroconsult in their latest report58 estimate a global figure of €1.4bn and a steady growth rate of
6–8% per annum. Projecting at this rate of growth from 2006 would lead us to estimate a European

market for VA services of around €420m, i.e. approximately one-third of the global market.

How much of this market has been built based on the marginal-cost data from ESA? Not very much,
according to the survey that EARSC carried out amongst its members on the occasion of a data policy
workshop held during the AGM. The VA companies are more often using high resolution data from
the commercial missions rather than the low or medium resolution data from the ESA missions
(although it should be recognised that the recent Envisat failure has taken away an important data
source for many).

However, there are some excellent examples of European companies making use of free data to
build or support their commercial business. Two of these are described below.

 BMT-Argoss, a Dutch company, has built its business in metocean services around data
coming from the US satellites: Landsat, Aquos and Terra, and from European satellites:
Envisat, ERS and Jason, each of which makes imaging observations and provides the data free
for online access. Many of the products and services offered are based on the free data
access and the company has built a successful business on this basis, supplying products such
as tidal heights and currents, wind and wave climate and bathymetry. Today this has resulted
in a business with revenues close to €10m and over 120 employees in the Netherlands and
UK. The CEO says that without free data, his business would look completely different.

57
The State and Health of the European and Canadian EO Service Industry; Vega, 2006.
58
Satellite-Based Earth Observation, Market Prospects to 2012. Euroconsult, October 2012.
42

 The second interesting example is Enviros, a Czech subsidiary of a UK company. It offers
services to investors in solar energy projects to provide an audit of the likely energy
production and hence revenue potential for a site. Enviros makes use of a database put
together originally by the Eumetsat SAF (the Satellite Applications Facility of the European
Organisation for the Exploitation of Meteorological Satellites) for climate change research. A
copy was purchased by the Joint Research Centre, the EC’s in-house science service, a few
years ago and re-processed into a database called PV-GIS (Photovoltaic Geographical
Information System), providing information on solar illumination levels at the Earth’s surface.
The database was developed as a support tool for EU policy makers – so similar to GMES in
many ways. The PV-GIS database was subsequently made available for external users on a
free and open basis and has been picked up and used by a number of companies. In the
period 2009 to 2011, Enviros has developed revenues of about €1.5m compared to annual
turnover of €3.5m. Other companies have also used this database as the basis for business;
for example, Geomodel (Slovakia) offer a range of consultancy services based on a derivative
of PV-GIS and A2-C (Switzerland) offer a software tool called Evalo which is sold or licensed
to companies as a promotional tool.

Limited but growing evidence
Although limited in size and evidential value, the cases appear to signal that freeing up EO data can
lead to market development and business growth. It is certainly clear that the interest in the data is
extremely high. As we saw above, the characteristics of GMES data, the market conditions and the
policy setting are all highly favourable to reap fully the PSI re-use potential. This raises an obvious set
of questions, such as: how high is this potential, what would it cost and when would it kick in. We
shall examine this in more detail in the next paragraph.

4.3 Economic effects of free re-use of GMES data – an educated guess

As seen in the previous paragraph, clearly, from a quantitative and statistical point of view, no
reliable extrapolations can be made where solid figures are lacking, in particular within the GMES
domain. Nevertheless, we can certainly try to come up with an educated guess as to the basic
developments, as we are able to compare the effects coming from the cases described above in
paragraph 4.2 with the more solid ones reported upon in chapter 3.

This being said, what are the basic developments?

PSB ‘losses’ much lower
First and foremost, and rather uniquely, the GMES re-use is not inhibited by any legacy. Thus, it can
do things right, right from the start. Where other PSB domains suffer from being bogged down in the
‘own-re-use, income-reliance trap’ – preventing them from moving to free re-use models, as this
would cause their immediate bankruptcy – GMES public data holders are not dependent on income
from charging users (either from the public or private sector). Consequently, costs incurred or
foregone incomes are relatively much lower, as they only comprise the costs for facilitating the re-
use. Accordingly, the PSBs’ accrued losses (represented by the plane in yellow) are much lower.

SMEs decisive for market development
Although less spectacular than in some other PSI domains (as in meteo), demand increases shown by
the Landsat and the Dutch and Swedish NSO cases indicate an exponential surge in volumes as well
as numbers of users.

Looking more closely at the character of the new users, it is very likely that, even to a higher extent
than in other PSI domains, SMEs will be highly represented, as the EO services sector is dominated by
SMEs. The last industry survey in 2006 showed less than 9% of companies in this domain have more
than 60 employees. EARSC is about to conduct a new survey and has identified over 250 companies

43

believed to be active in the sector. Today, no more than five or six companies in Europe would have
more than 250 employees or revenues over €50m, the cut-off definition for an SME under
international rules, making it likely that at least 90% of the companies active in this domain qualify as
SMEs59.

This is important as many studies have shown that it is through SMEs that the majority of new jobs
are created. For example Kane 201060 shows that start-ups in the US are responsible for creating
most net new jobs. Further, Koski61 shows empirically that in certain countries and certain sectors
(architecture and engineering firms) the availability of marginal cost geographic information (GI)
increases growth by 15% per annum compared to countries which do not make their PSI available for
re-use.

Low entry barriers sparks innovation
Technical universities throughout Europe establish a strong knowledge infrastructure, where many
future entrepreneurs are currently conducting academic and applied research. On opening up GMES
data, this nursery of talent will create an interesting cocktail of capable people and market
opportunities. This will particularly apply to applications of GMES data where the free availability of
data will attract high potentials, as the entry barriers are relatively low: brains, will power, a
computer and data (almost) come for free, and mobile devices allow for low-cost distribution of
digital information.

High dynamism and value feature steep curves
Put together, low entry barriers and high concentration of SMEs explain the movement of the purple
curve, displaying the benefits of the private sector: demand will increase rapidly in the beginning, but
will soon level out to a steady growth. Market dynamism effects will kick-in at the end of the growing
phase where successful start-ups are maturing and new staff are being hired.

This is also reflected in the steep rise of the private sector benefits in the harvesting phase, where
successful start-ups will flourish swiftly and are likely to be acquired by the big players on the basis of
potential rather than current performance, boosting economic value and allowing for an attractive
way out for the founders, often allowing them to become start-ups again after an earn-out period.
This is mirrored by the ‘second sowing phase’ and the cyclic shape of the curves.

Speed of effects
Another important difference with the other PSI domains is the expected speed of effects. It seems
likely that the phases are shorter, where the subsequent effects are strongly inter-related and follow
each other quite swiftly. Again, this is a result of low entry barriers and people willing to take risks,
relying on lessons learned from trial and error rather than neatly planned and tested campaigns.
Where in other PSI domains time lags in effects have been reported amounting to one or two years
(Pollock and Koski – see Annex 4), it seems reasonable to assume that for EO data this will be shorter,
possibly (0.5 – 1 year). Accordingly, one can expect the first real economic benefits to materialize in
the growing phase after a short time.

Taxation effects start early but slowly
Obviously, where the costs of the PSBs is limited to the re-use facilitation costs, not having to cover
up for the lost revenues resulting from free data re-use, the positive taxation effects kick in much
earlier, requiring almost no transitory funding. After a relatively slow growth, the returns for the

59
Whilst the large majority of EO service companies are SMEs in terms of employee and revenue numbers,
they do not qualify as SME’s under European rules as they are often owned by larger companies. However, the
nature of the sector means that they operate with a large degree of independence.
60
The Importance of Start-ups in Job Creation and Job Destruction, Tim Kane, Kauffman Foundation, July 2010
61
Does marginal cost pricing of PSI spur firm growth?; Heli Koski, ETLA, September 2011.
44

treasury go up steeply (presented by the green plane) as VAT, profit taxes and social security taxes
flow in.

Putting everything together, Figure 4-4 depicts the likely effects for the GMES domain. Obviously, it
should be kept in mind that this is a guesstimate and, it would require more in-depth measurements
– going beyond the objectives of this study – to acquire solid estimates on the economic effects likely
to kick in following an unconditional and long-term adoption of a free data re-use policy in the GMES
domain.

Non-economic
benefits
Phase 1 Phase 2 Growing Phase 3 Phase 4
Benefits
Sowing Harvesting Sowing 2 Benefits
created in private sector
• Business effect
euros
• Demand re-users • Employment
Additional tax
effect re- • Market effect incomes
users dynamism • Welfare effect
effect Income
Treasury

0 Time
X1 X2 ∞
Costs/missed
revenues

public sector
Costs • Income • Efficiency effect • Taxation
incurred effect PSB effect
in euros PSB

Figure 4-4: Chart of guesstimated benefits in the GMES domain

4.4 Wrap up
Hard facts and voluminous figures on the economic effects of freeing re-use of EO data are missing.
However, based on the few cases reported, there are good reasons to assume that the effects seen
in other PSI domain, as described in this chapter, will emerge from the free provision of GMES data.

Indeed, there are good grounds to believe that the specific characteristics of the GMES domain (type
of PSI, type of market), will turn data from the GMES into a show case for free re-use, as it is not
hampered by any financing legacies and combines a powerful combination of a strong academic
knowledge base, low entry barriers, a large potential of entrepreneurs and low-risk high-return
market circumstances. The eggs laid by the GMES “goose” may indeed be golden.

Of course the next question is: how to realize this dormant potential? That is addressed in the last
chapter, which holds our conclusions and recommendations.

45

5 THE CONSEQUENCES – THE WAY AHEAD, HOW TO MAKE IT WORK?

5.1 Introduction

Having examined the evidence from the other domains and compared this to the EO domain, there
are a number of lessons and conclusions that can be drawn for GMES data. We therefore take a look
at some of the implications of a free and open data policy, which feed in to the recommendations
made in paragraph 5.3.

5.2 Further considerations

A free and open policy for GMES data and information will offer the best possibility to grow the
market for EO services. Nevertheless, there are some possible implications.

Risk to private operators
For commercial data suppliers, there is no doubt that free data coming from the Sentinels will have
an impact on their business models. As we have seen, the resolution of the imagery from the
Sentinels is medium resolution (10m and above) whilst the commercial operators are mainly offering
a better resolution (better than 5m). There is, however, some overlap with some suppliers and also a
risk that some users will accept an information product with a slightly lower resolution if it is cheaper
than one based on commercial imagery. Hence, these two aspects represent a threat to some
operators.

This threat is somewhat counterbalanced. Firstly, the expectation is that access to the GMES data will

lead to an increase in the total market for EO services. The use of Sentinel data should also stimulate
the use of commercial data; greater awareness should ensure that the whole market increases in
size, allowing the incumbent players to piggyback on this beneficial development. Secondly, some
GMES services will require higher resolution data. In this respect, the commercial missions become
contributing missions and data will be purchased to meet GMES needs. If the data license holders
acquire full re-use rights they would be able to allow full re-use of this acquired PSI. More realistically
If the budget is insufficient such that only partial re-use rights are obtained, the residual rights must
be dealt with appropriately. Other measures may be foreseen to help limit the impact on more
vulnerable operators.

PSI for global or European re-use?
Then there is the question of whether the GMES data and information should be made available
globally or just for European users? Some countries, notably the US and Brazil, have taken the
decision to allow anyone to simply register and have access to the data. Data and information
sharing within GEOSS (Global Earth Observation System of Systems) will further encourage this trend.
The EU, as a major international partner, wishes to take the lead by making data freely available on a
global basis.

Where European taxpayers’ money has been relied upon to produce the GMES data and where the
financing from the State budget is aimed at creating economic and social benefits within Europe
through advantageous access and re-use schemes, these (competitive) advantages would be watered
down if non-European entities were allowed the same terms. Of course, such exercise of IPRs creates
red tape and overheads, whereas, from a purely economic perspective – as we saw in chapter 2 –
provision of data against marginal (or no) costs would make most sense. However, from an EU trade
policy perspective, bargaining would lead to an agreement that is mutually and reciprocally

46

beneficial, allowing any world citizen to (freely) re-use satellite data, including those produced by
non-EU countries.

Furthermore, it must be remembered that Europe and the US have pursued a policy of public-private
partnership. As discussed above, commercial firms have invested alone or alongside national
governments to develop and launch high resolution satellite systems. We recommend that a clear
understanding should be developed regarding the specification of any future public systems so as not
to undermine any of the commercial business models. Such an agreement needs to be recognised
internationally as well or there could be more confusion in the commercial market, further inhibiting
any future private investment.

Thus there is a clear and highly important role for the European Commission to promote the broad
European interests coming from GMES. Firstly, using diplomatic negotiations to ensure that the value
of the data being made available, paid for by the European taxpayer, is exploited to the maximum
and to be certain that other countries using the data are fully cognisant of the value of the resource
to which they are being given access. Secondly, to develop real economic value through supporting
European industry to gain access to markets arising from the investments being made.

Are the business models fixed?
Whilst the public sector has a responsibility to ensure that the information is available, it is not
always charged with producing the information itself. Much PSI generation is outsourced, and it is
important not to confuse these two roles. We showed earlier that some boundaries change - in both
directions. For TomTom, supplier of location and navigation products and services, the boundary for
generating and distributing traffic information has moved from within the public sector into the
private sector. For Landsat, the operations of the satellite and the geo-correction of imagery have
moved from the private sector back into the public sector.

That the boundaries can move is healthy and should be reviewed on a regular basis to see if a better
optimisation is possible. However, there should be some agreement regarding changes and for
maximum commercial exploitation it is necessary that these changes are planned and transparent in
order that investment is not discouraged. This necessitates a close and sustained dialogue between
public and private representatives.

International organisations
As an EU programme, several international organisations have a role to play in an operational system
for GMES. ESA and Eumetsat will operate the Sentinel satellites whilst ECMWF (European Centre for
Medium-Range Weather Forecasts) will be a leading player in the atmosphere and climate services.
All of these agencies are formed under international rather than EU treaties which has implications
when it comes to implementing a coherent data policy.

In other information services, the European Environment Agency (EEA) may take the lead for the
land monitoring task, EMSA has a principle role to play in some of the marine and security services,
and the EU satellite centre could lead other security services.

5.3 Policy implications

Towards a clearly defined and harmonized European data re-use policy for GMES
We have shown that the data and information coming from GMES Sentinels are collected and paid
for by the public sector in the process of exercising its public task. Accordingly, it clearly qualifies as
PSI, in fact as re-useable PSI par excellence. Therefore, it should be recognised as such by the EC in
the context of the on-going review of the PSI Directive. This would then also be recognised in the
GMES data and information policy to be in-line with the PSI Directive.

47

The PSI nature of other data and information gathered under GMES shall also need to be taken into
account according to the different types:

− GMES Sentinel data as PSI fully paid for and owned by the public sector
− GMES contributing missions with residual rights according to the data licences agreed with
each supplier

Several European agencies operating under international treaties (non-EU) are concerned in the
GMES programme. They should follow the overall policy applied to the GMES data and information.
The EC will need to clarify the roles of all GMES partners and monitor their application.

Recommendation 1: The nature and potential of GMES data and information should be recognised
within the review of the PSI Directive and within the GMES data policy to be adopted in the near
future for the different data and information types.

Towards clear demarcation of public task responsibilities and private sector market domain
In other PSI domains the presence of PSBs selling in the open market has inhibited the business of
commercial operations. The most marked example is meteorology. In the EO/GMES domain the PSBs
have less history in the commercial market even though they sometimes compete with the private
sector.

Successful exploitation requires the development of products and services that can be taken into
other markets. The necessary competences must be developed and remain within industry.
European PSBs should focus on the use of the products and work together with industry to ensure

that the home market is a reference for exploitation in exports. The market for PSI re-use has been
successful in the US where the roles of public bodies are more clearly defined. Meteorology provides
the best example.

At the same time, many of the GMES services are generated in teams comprising a mix of both public
and private entities. There are signs that public entities seek to maintain their roles despite services
moving from research to operational status. This could be because the institutes concerned fear a
lack of long-term, stable research funds and seek to secure their future through operational funds.

In all areas of GMES the roles and responsibilities of the public and private sectors need to be
addressed. There should be well-defined boundaries between public and private tasks. If industry is
to invest in product or market development it can only do so with clarity over the boundaries
between public and private sector roles. Whilst public-private working together is good, this
relationship must be built upon the appropriate relationships where both parties can see a stable
future.

The key question is where the competences and key tools are to be developed and maintained.
Where necessary skills exist in the public sector, these can best be used by industry if there is an
agreement to a long-term commitment that they will be sustained. Otherwise the goal should be for
PSBs acting as agents of government with services procured from the private sector. This would allow
the PSB to allow re-use of the information whilst industry has the skills, competences and capacity to
exploit the knowledge and data in other markets.

Recommendation 2: Ensure clarity of the role of public sector bodies within GMES based on a long-
term and stable arrangement.

48

Towards an open and responsive governance model
The governance of GMES is still an open issue. Industry currently has no formal voice in the
governance of GMES. As new products and services are requested by the public sector, the way in
which these are brought to market should be agreed by all parties. The public sector may wish to
have a new service developed and run under a GMES brand. This may be the appropriate way to
engage industry but companies may also prefer to invest themselves in order to gain a better
position in the market (faster entry, return on investment, etc.).

A stable, transparent and clear governance of GMES is necessary to ensure the appropriate long-
term conditions are provided to underpin investment decisions. Such a structure should also act to
ensure that the needs of European public sector users are met by federating demand, ensuring
quality of service and maintaining the necessary long-term relationships between other PSBs and
industry.

Recommendation 3: The dialogue between public and private sectors should be fortified to ensure
that industry’s voice is heard. As a key player, industry must have a voice in the governance of GMES.

Towards mitigation of collateral damage to private business models
The availability of free GMES Sentinel data will have an impact on the market for EO data coming
from existing private operators even though these mainly provide data at higher resolutions than
that of the Sentinels. Some substitution of data can be expected as some users will choose to accept
a poorer product (i.e. lower resolution) but at lower cost (based on zero cost data). This will be offset
by two mechanisms: the increase in the overall size of the EO market and direct sales for GMES
products. Nevertheless, some care must be taken to ensure that commercial business models are not
undermined as a result of the public activity. This must include an understood and agreed

demarcation between the specification (notably resolution) for future public missions including those
that are subject to international agreements (see recommendation 8).

Recommendation 4: Ensure adequate measures are taken to protect the investments of private
companies in commercial satellite systems.

Towards maximizing synergies between GMES data and European R&D policies
GMES provides an excellent opportunity to develop a world-leading European industry. The large
volumes of data will open the possibility for new products and services. By lowering the entry
barriers to new companies, a free data policy will stimulate innovation and creativity through the
value chain. It is therefore essential that Europe continues to invest in research and development
(R&D) both to ensure that this innovation and creativity can be applied to public-sector needs and to
provide the industry with the means to address globally issues that lead to export success. With
appropriate public sector support, there is no fundamental reason why Europe’s industry cannot take
a strong lead in this business area.

Recommendation 5: Ensure adequate R&D funding is available for industry at both the national and
European levels.

Towards catalysing the full benefits of SME potential
The EO services sector, which will be the primary motor for the exploitation of GMES, is dominated
by SMEs complemented by a few large players. This makes for a very dynamic and exciting sector
where entry barriers are very low and researchers or entrepreneurs can easily establish a new

49

product or service. At the same time, the few larger players provide channels that can help exploit
new export opportunities and that also provide easy exit opportunities if the business idea becomes
successful. If tackled correctly, the stimulus of a micro-environment of start-ups and young SMEs as a
consequence of a specific public programme could become a showcase for good public policy.

Recommendation 6: In a dialogue with industry, look for specific measures that can be taken to help
develop and sustain start-ups and SMEs in the EO services sector.

Towards common standards, driving economic growth in Europe and beyond
A proven way to help European industry establish itself as a global leader is to establish a set of
product standards that can become accepted world-wide. In addition, the use of these products in
the European “home” market is an essential reference for selling elsewhere.

Standards exist for data formats and for processes, and Inspire provides a European framework for
many geographic information products. But today no certification scheme or product standards exist
for the products that are generated by the EO geo-information services industries. As the industry
moves from being R&D-based to an operational one, some of the more commonly generated
products could be defined in a standard way. Indeed, in certain commercial sectors, customers are
starting to call for standards for certain operational products.

Since free data lowers the entry barriers for everyone, there is a risk of low quality products being
made available; this can damage the market. Consequently, a scheme to help European companies
develop a quality image and a professional industry should be encouraged.

Recommendation 7: European policy makers should work with industry to establish a comprehensive
and recognised set of standards for GMES products and services.

Towards a reciprocal external data policy
Should the GMES data and information only be free to European users or should the policy be
extended globally? The issue provokes strongly differing opinions. Whichever policy is adopted, the
goal should be to develop maximum economic benefit for the European taxpayer.

In the domain of navigation services, free access to GPS signals, excepting specialised services, e.g.
security, has been the favoured policy from the outset. The result has been a rapidly expanding
market for downstream services which has been exploited by industry throughout the world. Today,
the market for geo-location services is estimated at around €90bn.

If a free and open data policy will help develop the market for EO services in Europe this is equally
true in other parts of the world. GMES Sentinel data is global in nature and could be made available
to global users hence promoting the image of Europe as a global leader. However, this could enable
non-European companies to gain a competitive advantage unless European industry (and indeed
public institutions) can also have free access to data coming from other national or regional satellite
systems.

Hence, reciprocity is sought with other national satellite operators backed up by two measures:

1. A framework of international agreements that ensure reciprocal access to international
datasets and that include a limitation on the performance of future satellites included in the
policy.
2. Technical controls that can be applied if other nations abuse the free data policy.

50

Recommendation 8: If the EC and ESA make GMES Sentinel data freely available to global users, full
access to international datasets must be assured on a reciprocal basis for European users through
international agreements and backed by technical security barriers to control access.

Towards maximising the economic returns
Industry is the motor by which economic returns can be realised. Many of the economic benefits
flow into the public sector and industry needs to learn how to capture these and apply the
knowledge into other markets; especially exporting outside of Europe.

To achieve this requires a special form of public-private partnership whereby the public sector and
industry work together to develop, deliver and sustain new information services. Industry needs to
be able to understand the mechanisms and drivers whereby geo-spatial information is turned into
societal value. The public sector needs to understand the issues faced by the European industry in a
global competition and to be ready to help overcome the international barriers to business. GMES
provides an excellent opportunity to develop this partnership given the very strong public role in its
gestation, development, operations and most importantly its exploitation.

Recommendation 9: Industry and the public sector to work together to establish a new partnership to
ensure that the fruits of GMES public investments lead to a world-beating European industry.

5.4 Final observations

This study has brought together the two different worlds of PSI re-use and GMES and it seems that
each has something to learn from the other.

GMES has a long history of gestation. Since the Baveno Manifesto first described the needs
of the EU, the idea has developed considerably but has yet to become reality. The first
satellites are almost ready for launch but continued discussions over funding are still
delaying the entry into service. This study will hopefully help to bring some additional
support by showing how by applying the economic laws and legal principles surrounding PSI
re-use the data and information that will be generated in the GMES domain to meet public
needs can also lead to substantial socio-economic benefits in due course at low cost.

PSI re-use is a policy to free up public data collected by governments in the course of fulfilling
their duties. The policy has recognised neither GMES nor EO data more widely as coming
under its scope. The new Directive planned for 2013 could be an opportunity to change this.

GMES seems to be a perfect type of PSI where there is a clear subsequent second value: it serves the
public needs in the framework of the public task and it allows for re-users, SMEs with potential to
create innovative new products and new jobs. The evidence coming from other domains suggests
that the lack of legacy organisations and the dynamics of the industry offer good prospects for new
products and services to be developed. In fact, GMES data appears to represent one of the strongest
business cases for a free re-use policy, where returns for the private sector (innovation and increased
turnover), the public sector (more efficiency, higher tax returns and more effective policy making)
and the European citizen (more employment, better services, better quality of life) are so eminent.

The evidence, albeit being short on quantitative data, when put into the context of evidence from
other PSI domains, appears to be quite convincing and provides us with a number of lessons learned.
These are translated into a number of policy recommendations of a fairly high level of aggregation.

51

Obviously, where more detailed and specific (market) knowledge would allow for more tailor-made
decision making (in terms of scope, focus, phasing, also taking into account the supra EU dimension),
there appears to be a strong business case for further economic research to beef up the quantitative
basis. This would have the additional benefit of providing a sound base for measuring the success of
the policy in the future.

Furthermore, the various policy recommendations that are made will need to be expanded and
worked upon. Responsibility for implementing them, if they are accepted by the broad community
and especially the political decision makers, will fall with various bodies and institutions. We look
forward to working wherever necessary to refine, define and implement those policies that can help
the industry in Europe develop into a first class supplier for the public sector and an excellent and
successful exploitation of GMES data and information in other markets.

We now look forward to presenting the results of this study to political influencers and decision
makers coming from both communities; PSI re-use and GMES. In so doing, we hope to re-enforce the
messages concerning the benefits that GMES can provide to Europe, its industry and all its citizens!


52

ANNEX 1 – ABOUT THE AUTHORS

Geoff Sawyer, BSc (Electronics), MBA
Geoff has held senior management positions in the space industry and numerous
representative positions in the UK and Europe. Geoff was the radar systems
engineer responsible for the ERS-1 synthetic aperture radar and after many steps
was, until recently, EADS Vice President Corporate Strategist for Space. In
addition to his extensive industrial experience, Geoff spent three years working
for the European Commission where he was responsible for supporting the
creation of the GMES initiative. Geoff is very well known in the space and earth
observation sectors and brings his depth of knowledge and wealth of experience
as Secretary General of EARSC to support the ambitions of the geo-information
industry that EARSC represents.
geoff.sawyer@earsc.org

Marc de Vries, BSc EC, LLM
Marc has professional degrees in both law and economics (Utrecht 1991). He has
been active in the field of PSI re-use for more than 15 years, both at the national
and European levels. In 2001 he started his own business under the name
‘Citadel Consulting’. Having served clients in the public and private sectors in the
Netherlands and beyond (EC institutions in particular), Marc has a wide network
of contacts and is frequently invited to speak at conferences as a subject matter
expert. He has published various books and articles on PSI, highlighting the legal,
economic and policy perspectives.
info@devriesmarc.nl


53

ANNEX 2 – ABBREVIATIONS AND SEMANTICS

Throughout this study numerous key terms are used. Grasping their exact meaning from the outset is
essential. Obviously, where possible, we rely on existing and accepted definitions from authoritative
sources. The key terms (in alphabetical order):

Term Meaning (and reference to source where possible)
Data All data that is generated by a PSB directly from exercising its public
task. Thus, any data (and its value) that is added outside the framework
of the public task is excluded.
Earth Observation The term is generally applied to non-defence surveillance of the Earth’s
(EO) surface using satellites and other sensors. It is used to describe the
monitoring of the natural and built environment and to provide
information leading to human intervention.
EO Data Data coming directly from the satellites or other sensors.
Global Monitoring for A European programme to develop a system including satellites to
Environment and collect information necessary for European policy makers to take
Security (GMES) informed decisions. The EO data collected will be turned into
information on environment and climate change, natural resource
management and security of the citizens in particular mitigation of the
impacts of natural disasters as well as many other policy areas.
EO Information The product of subjecting EO Data to processing often in combination
with data from other sources e.g. in-situ data.
PSI charging model: “Setting a price equal to average long-run costs (including, for example,
Cost recovery all fixed costs related to data production).”
(‘The Economics of Public Sector Information’, Cambridge Report 2008)

PSI charging model: “Setting a price equal to the marginal cost of supplying data (that is,
Marginal cost simply the cost of actually transmitting the data to someone).”

The digital nature of many forms of PSI implies marginal costs of
approximately zero.
PSI charging model: Setting a price lower than average long-run costs and higher than the
Partial cost recovery marginal cost of supplying data.
Sub-category: Re-use facilitation cost recovery
Setting a price equal to average long-run re-use facilitation costs.

Re-use facilitation costs correspond to all additional costs incurred by a
PSB to enable and facilitate re-use of PSI. These costs notably include
costs for data transfer to re-users (such as servers), anonymization, data
re-formatting for re-users and re-user helpdesks. The collection and
processing of the data within the public task is not included in the re-
use facilitation costs.
PSI charging model: “Setting a price to maximize profit given the demand faced by the PSB.
Profit maximization Where the product being supplied does not face competition then this
will naturally result in monopoly pricing.”
PSI charging model: “Setting a price equal to zero.”
Zero cost (‘The Economics of Public Sector Information’, Cambridge Report 2008)

Public sector body “A State, regional or local authorities, bodies governed by public law
(PSB) and associations formed by one or several such authorities or one or

54

several such bodies governed by public law.” (Art. 2(1) Directive
2003/98/EC)
Public sector "[E]xisting documents (holding content, whatever its medium and any
information (PSI) part of such content) held by PSBs of the EU Member States." (Art. 1(1)
j-o 2(3) Directive 2003/98/EC)
Public task Setting the scope of the public task (and its financing) is a political
decision taken at the national level (not the European level).
Nevertheless, to determine the public task, some thumb rules apply:

PSI is produced under the public task if:
a. the PSI is the result of the legal regime under which the PSB works.
Example: all constitutions assign the task of producing court
decisions to national courts; hence, their case law is produced under
the public task and falls under the PSI Directive.

b. the production, processing, or distribution of the PSI falls under the
core business and responsibility of the PSB. Example: the sole reason
for setting up the Dutch Chamber of Commerce was to maintain the
Dutch business registers; hence, those registers are produced under
the public task and fall under the PSI Directive.

c. there is a strong public interest involved in the production,
processing, or distribution of the PSI concerned, whereby society at
large benefits (i.e. the benefits do not accrue to just a small group of
people). Example: maintaining the quality of cadastral data is key as,
otherwise, there would be even higher risks involved in buying
property (i.e. the buyer might risk paying money to a person other

than the real property owner). Therefore, producing cadastral
information is done under the public task and falls under the PSI
Directive.

d. without the engagement of the government, the PSI would not be
produced because of market failure, i.e. the market would not be
able or willing to perform this task. Example: the private sector
cannot afford to build and launch the weather satellites required to
gather meteorological data. Therefore, national meteorological
services undertake these activities, which are regarded as falling
under the public task. Thus, the output falls under the PSI Directive.

Re-use Any use of PSI outside the public task including use by the PSBs
themselves (including the PSB that has produced the PSI under its
public task).

“Use by persons or legal entities of documents held by public sector
bodies, for commercial or non-commercial purposes other than the
initial purpose within the public task for which the documents were
produced. Exchange of documents between public sector bodies purely
in pursuit of their public tasks does not constitute re-use.” (Art. 2(4)
Directive 2003/98/EC)
Sentinels A system of five satellites being developed by the European Space
Agency, which will be used to provide EO data for GMES.

55

ANNEX 3 – SNAPSHOTS OF THE ‘OLD’ AND ‘NEW’ PSI DIRECTIVE

A. The 2003 PSI Directive

The Directive stems from the Green Paper on PSI (1998) and a subsequent Communication (2001).
It was adopted on 17 November 2003 and published in the Official Journal L345/90 of 31 December
2003.

The Directive is addressed to all Member States and, under the principles of European law, imposes
the obligation to implement the rules therein ultimately by 1 July 2005. Although only four Member
States managed to meet the implementation deadline, all have now put the framework in place in
their national legislations. The Directive has four main objectives: (1) to stimulate the further
development of a European market for PSI-based services; (2) to enhance the cross-border use and
application of PSI in business processes, including publishing; (3) to fortify competition in the internal
market; and (4) to address divergence as to re-use rules between Member States (as this will impede
meeting the stated objectives). The Directive allows Member States to implement measures going
beyond its minimum standards, thus allowing for more extensive re-use (which we now see
happening in quite a few Member States).

The Directive consists of 25 recitals, and 5 chapters subdivided into 14 provisions. In essence, it
imposes obligations on Member States’ public sector bodies (PSBs) once they decide to allow re-use
of their PSI (Article 3) and provides corresponding rights to re-users thereof. Classic internal market
principles of fair competition and transparency are the main building blocks of the Directive. Article 2
holds definitions of the key concepts such as ‘documents’, ‘PSBs’ and ‘re-users’. Furthermore, it
connects to existing terms in other legal frameworks, such as ‘personal data’ and ‘services of general
economic interest’. Its area of application (Article 1) is mainly set by the term ‘public task’, under

which documents need to be supplied to be potentially subject to the re-use rules, and by some
exemptions related to the other interests connected to the PSI (e.g. absence of access rights or
presence of third-party intellectual property rights) or types of business of PSI holders (e.g. public
broadcasters, or educational, research and cultural establishments).

Once PSBs allow for re-use of their PSI, the Directive imposes obligations on how to deal with
requests for re-use and redress (Article 4), and the practicalities of making documents available, for
example, formats (Article 5) and accessibility (Article 9). Transparency of conditions of re-use and
tariffs are addressed in Articles 7, 8 and 6, the latter establishing an upper limit for charging, based
on costs incurred to produce the information, together with a reasonable return on investment.
However, lower charges (or no charges at all) can be applied and upon request, PSBs have to indicate
the calculation base for the charges. Where licences are imposed, they should never restrict
competition. This principle is elaborated upon in Article 10, which obliges PSBs to apply equal
conditions to comparable types of re-use, including their own commercial re-use activities (to avoid
cross-subsidies), and Article 11, which bans exclusive PSI re-use deals between PSBs and re-users
(with an exception for those necessary for the provision of a service in the public interest). Finally,
the Directive also provides for regular reviews of its application to be conducted. As foreseen in
Article 13, the first one was carried out in 2009, whereby the EC decided not to make any
amendments but rather to monitor the sound application of the Directive and to initiate a
subsequent review in 2011.

56

B. The 2011 Proposal for a new PSI Directive

Following the 2011 review, in December 2011, EC Commissioner Ms Neelie Kroes launched the
Digital Agenda for Europe and the Europe 2020 strategy that aims for smart, sustainable and
inclusive growth. The proposal for a new PSI Directive forms an essential part hereof.

The line of reasoning is that since 2003 — when the first PSI Directive was adopted — the amount of
data in the world, including public data, has exploded and we are witnessing a continuous revolution
in technologies. Accordingly, the rules adopted in 2003 can no longer keep pace with the rapid
changes. Therefore, the proposal seeks to provide the market with an optimal legal framework to
stimulate the digital content market for PSI-based products and services, including its cross-border
dimension, and to prevent distortions of competition on the EU market for the re-use of PSI, by
eliminating persistent and emerging differences between Member States in their exploitation of PSI,
which hamper realisation of the full economic potential of this resource.

The main relevant changes concern:

Access = re-use
Under the 2003 Directive, the decision of whether or not to authorise re-use remains with the
Member States or the public sector body concerned. The new proposal holds a clear obligation for
Member States to make all generally available documents re-usable.

Accessibility on an INSPIRED basis
To facilitate re-use, public sector bodies should make documents available through machine-
readable formats and together with their metadata, where possible and appropriate, in a format that
ensures interoperability. For example, by processing PSI in a way consistent with the principles
governing requirements under the INSPIRE Directive.

Charging against marginal costs as a rule
Where charges are made for the re-use of documents, they should in principle be limited to the
marginal costs incurred for their reproduction and dissemination, unless exceptionally justified
according to objective, transparent and verifiable criteria. This exception could apply to those public
sector bodies that would go out of business when implementing the marginal charges regime, as
they cover a substantial part of their operating costs from the exploitation of their PSI. The burden of
proving that charges are cost-oriented and within relevant limits should lie with the PSB.

PSI re-use watch dog brought in
Redress is being made easier as Member States will need to appoint an independent authority that is
vested with specific regulatory powers regarding the re-use of public sector information and to
whom re-users can turn in case of denial of requests for re-use. Its decisions are to be binding upon
the public sector body concerned.

Enhanced EC monitoring and guidance
Finally, the EC intends to ensure that the Member States will report on the extent of the re-use of
PSI, the conditions under which it is made available and the work of the independent authority.
Furthermore, it may give guidance to Member States as to implementation of the Directive in a
consistent manner, particularly on charging and calculation of costs, on recommended licensing
conditions and on formats, after consulting interested parties.

57

ANNEX 4 – STRUCTURED SUMMARIES OF CASE STUDIES
Trading funds study

Models of Public Sector Information
Provision via Trading Funds
Introduction This summary is based on the study ‘Models of Public Sector Information Provision via
Trading Funds’ by Newbery, Bently and Pollock, commissioned by the United Kingdom’s
Department for Business, Enterprise and Regulatory Reform and HM Treasury. The study
uses mathematical analysis.

Context − The study provides estimates for the costs and benefits of marginal cost pricing in
relation to bulk, digital PSI data from the Met Office, Ordnance Survey, the UK
Hydrographic Office, the Land Registry, Companies House and the Driver Vehicle
Licensing Agency (DVLA).
− The study uses prior experiences of agencies adopting marginal cost pricing.

Key findings − Elasticity of demand varies depending on the PSI, but for the products associated with
the PSI, average elasticity was estimated at between 1 and 2.
− The average time delay for the effects to take place is 1.5 years.
− Overall welfare increases when marginal cost pricing models are used:
o By Companies House for CD ROM and bulk data and image products: estimated
net welfare gain is around €2.3m.
o By the Met Office for wholesale products net welfare gain of €1.3 m, with
investment returns of approximately 500%.
o By Ordnance Survey for large-scale topographic and transport network products:
estimated net welfare gain is around €194m. Estimated £85m cost to government

would be reduced by increased tax income and introduction of a subsidy to spread
costs across departments.
o By the Hydrological Office for digital UK charts, digital publications and licensing
products: estimated net welfare gain is around €429,000.
o By the Land Registry for property price data and polygon geographic information
system data: estimated net welfare gain is €1.5m. Due to the limited amount of
information, no figures were estimated for bulk data downloads, but significant
overall welfare gains were estimated.
o By the DVLA for anonymised, bulk data and mileage data: estimated net welfare
gain is around €4.6m. The relative inelasticity of demand for registrations may
permit financing from increased registration fees.
− Registration-based entities may be able to make up revenue losses through changes
to registration charging policy. Non-registration-based entities would require
additional government funding.
− The establishment of an adequate regulatory regime is of crucial importance.

PSI domain Company register, land registry, meteorological, hydrological, vehicle and ordnance survey
data.

Years of data Evidence for mathematical analysis taken from sources from 1996 onwards.

Year published 2008

Source http://www.berr.gov.uk/files/file45136.pdf

58

Pollock study

The Economics of Public Sector
Information
Introduction This summary is based on the study ‘The Economics of Public Sector Information’
by Rufus Pollock. It uses mathematical analysis to assess who should finance PSI
re-use and the regulatory structure needed.

Context − The study focuses on digital, non-personal and upstream PSI. The types of
data considered are large and coherent datasets rather than individual pieces
of data. It proceeds on the basis that marginal costs are zero when the
information is digital. Transaction and transition costs are not included in the
welfare analysis.
− Elasticity of demand estimates are drawn from national experiences of
improving re-use conditions.
− When assessing welfare, the study takes account of the fact that PSI is often
sold to intermediaries rather than direct consumers and the positive effect of
new products and services.

Key findings − The evidence on price elasticity is limited, and its value will be determined by
the nature of the product at issue. Nevertheless, the study estimates that
elasticity is generally greater than 1, and the range for the kinds of products
that are the subject of this study is between 0.5 and 2.5.
− Given that the estimated elasticity of demand for digital, upstream data is
greater than 1, marginal cost pricing is found to be preferable.
− Good reasons for pricing at marginal cost or below include the following:

o Average costs distort the market because there is usually a high mark-up
to cover costs.
o Demand for digital PSI is likely to be high, and growing and average costs
may hinder innovation.
o Competition increases as public sector monopoly on data is removed.
o Governments are already providing large contributions to fixed costs; the
public should enjoy the benefits of that contribution.
− For those legally required to update registers (‘updaters’), rather than ‘users’
of PSI, an above marginal costs pricing model is preferable: elasticity of
demand is less than 1 and there are no systematic distortions of ‘updaters’’
willingness to pay.
− To prevent market abuses and regular funding changes, public bodies in
control of PSI should be regulated transparently and independently.

PSI domain Digital, non-personal and upstream PSI.

Years of Mathematical analysis is based on evidence from a wide range of previous (peer)
measurements studies.

Year of 2009
publication

Source http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe0920.pdf

59

Pettifer 1 study

PSI in European Meteorology – an
Unfulfilled Potential
Introduction This summary is based on the study ‘PSI in European Meteorology – an Unfulfilled
Potential’ by Richard Pettifer, Chairman of the Association of Private
Meteorological Services (PRIMET).

Context − The study assesses the way in which meteorological PSI is exploited in
Europe, and compares the estimated market and growth in the US and EU.
− It proceeds on the basis that, in general, meteorological PSI is available on a
cost-recovery basis in Europe and on marginal or zero cost bases in the US.

Key findings − Cost-recovery pricing in Europe may be distorting the market:
o National Meteorological and Hydrological Services (NMHS) are monopoly
wholesalers to their competitors on the value-added products market;
o NMHS may restrict re-use and adopt complex licensing systems; and
o cost-recovery pricing may raise data costs to such an extent that
commercial re-use is not economically feasible.
− If PSBs must maximise their PSI sales revenue, they may have to employ
complex licensing regimes that are expensive to manage.
− Where NMHS provide free (often low quality) value-added products,
opportunities for low-end commercial re-use are low, particularly for SMEs.
− The European market in value-added meteorological products is estimated
around €530m per year; the end-user market for weather related services is
estimated around €2 x 1011 per year. Therefore, only 0.3% of the European

market is being supplied, compared to estimates of 0.7% in the US.
− Since 2006, overall market growth in Europe has been less than 2%, with:
o the 28% private sector share growing at around 25% per year; and
o the 75% NMHS share declining by around 1.5% per year.
− If meteorological PSI was provided on a marginal cost basis throughout
Europe and the market became as well penetrated as that in the US, then the
actual market size would be around €1,390 x 106 per year; net revenue to EU
central treasuries would increase by around € 290 x 106 per year.
− Moving forward, the study suggests that:
o At the least, there should be complete, clear and public accounting
between the commercial arms of NMHS and its parent agency, and that, at
best, there should be complete technical and economic separation.
o States should be involved with regulating both the quality of PSI and its re-
use on the market, particularly due to the safety implications.

PSI domain Meteorological information

Years of Relies on data from various studies published from 2002
measurements

Year of 2011
publication

Source http://www.primet.org/documents-mainmenu-29/424-psi-in-european-
meteorology-an-unfulfilled-potential

60

POPSIS study

Pricing of Public Sector
Information Study

Introduction The ‘Pricing of Public Sector Information Study’ (POPSIS study) was prepared by
Deloitte and others under an assignment from the European Commission. It
assesses models of supply and charging for PSI and implications of price changes.

Context This study assesses different models of supply and charging for PSI and their
effects through the analysis of 21 case studies. The cases cover a wide range of
PSBs and different PSI sectors (meteorological data, geographical data, business
registries and others) across Europe. It examines the charging practices, ranging
from zero and marginal cost models to partial and full cost-recovery regimes. The
analysis focuses on the effects of PSI charging models on the downstream
market, PSI re-users, end-users and impacts on the PSB itself.

Key findings − The case studies show a clear trend towards lowering charges and/or
facilitating re-use (16 out of the 21 cases).
− Where cost-recovery regimes are applied, the calculation basis for setting PSI
re-use charges is weak, oriented towards filling budgetary gaps.
− PSI re-use revenues of PSBs range from relatively small to extremely small
when compared to the total budget of the PSB concerned. In half of the
cases, these revenues constitute less than 1% of the PSB’s entire budget.
− Based on their own raw data, the number of PSBs that exploit added-value
products is limited (seven out of 21 cases) and appears to be decreasing.
− In those cases where PSBs moved to marginal and zero cost charging or cost

recovery that is limited to re-use facilitation costs only:
o The number of re-users increased by between 1,000% and 10,000%,
attracting new types of re-users, in particular SMEs.
o Demand volumes expand strongly (up to 7,000%), where costs hardly
increase. Once re-use facilitation processes are properly organized, they
become sub-routines within the PSB.
o In some cases, PSI sales revenues can remain stable or even increase
after drastic price cuts due to the growing demand.
o Zero cost pricing has the additional advantage that transaction costs
decrease significantly. This decrease applies not only to administrative
costs, such as invoicing, but also to costs related to the monitoring of
compliance with licensing arrangements.
o Intensified ties PSB-re-users create better data quality and efficiency.

PSI domain General PSI

Years of 2006–2011
measurements

Year of 2011
publication

Source http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012/
models.pdf

61

Koski study

Does Marginal Cost Pricing of Public
Sector Information Spur Firm Growth?
Introduction This summary is based on the study ‘Does Marginal Cost Pricing of Public Sector
Information Spur Firm Growth?’ by Dr. H. Koski of ETLA – The Research Institute
of the Finnish Economy.

Context − The study assessed 14,000 firms in the architectural, engineering and related
technical consultancy sectors, located in 15 countries62, some operating
maximum marginal cost pricing models, others cost-recovery pricing models.
− Industries in these sectors make significant use of geographical information,
and may also supply services requiring geographical information, e.g. digital
mapping, navigation and map data solutions.
− The study analysed the effect of maximum marginal cost pricing for
geographical PSI on the firms’ growth performance, measured by reference
to the firms’ real sales growth over a particular time period, taking into
account the effect of the availability of geographical PSI (e.g. the use of web
portals), the price of such data, and whether it was gathered and managed by
a nationally responsible department or body.
− Changes in national PSI policy included:
o The reduction of prices of digital cadastral maps by up to 97% (Austria).
o A free web portal for essential cadastral geographic data (Spain).
o A free web portal for spatial data (Australia).

Key findings − The average annual growth rate was about 15% higher amongst firms located

in countries with maximum marginal cost pricing for geographical PSI
compared to those operating on a full or partial cost-recovery basis.
− Changing to marginal cost pricing had the greatest impact on SMEs, as:
o it substantially contributed to the sales growth of SMEs, but did not have
a significant impact on the sales growth of larger firms; and
o SMEs increased sales growth by 7% on average after one year and 19% on
average after two years.

PSI domain Geographic information in the architectural, engineering and related technical
consultancy sectors.

Years of 2000–2007 where possible; 2003–2007 otherwise.
measurements

Year of 2011
publication

Source http://www.etla.fi/files/2696_no_1260.pdf

62
Australia, Austria, Czech Republic, Denmark, Finland, France, Germany, Italy, the Netherlands, Norway,
Poland, Portugal, Spain, Sweden, the United Kingdom and the United States of America.

62

Spanish cadaster case

The Spanish Cadaster

Introduction This case was spotlighted in a Finnish study (Does Marginal Cost Pricing of
Public Sector Information Spur Firm Growth? by Heli Koski, ETLA) that assessed
14,000 firms in the architectural, engineering and related technical consultancy
sectors, located in 15 countries, analysing the effect of maximum marginal cost
pricing for geographical PSI on the firms’ growth performance. Further
information on this case has been drawn from the European Commission’s
2011 POPSIS study (Pricing of Public Sector Information).

Context − Since 2003, the Spanish Oficina del Catastro (the Spanish Cadastre/Land
Registry) has put increasing amounts of geographical data online and, from
2010, has facilitated electronic land registry certification.
− From June 2004, free access to cadastral maps by non-commercial users
was provided through the online portal IDEE (in English, the ‘Spatial Data
Infrastructure Portal’). In April 2011 free access was also extended to
commercial re-users, and a new model allowed mass downloads.
− The Spanish Cadastre runs regular training sessions for users of the portal,
to increase awareness and obtain feedback.

Key findings − From 2004–05, cartography data consultations increased by about 700%. In
2010, the increase compared to 2004 was over 25-fold.
− From a cost-benefit analysis, estimates for savings to the Spanish taxpayer
from the online access and digital certification range from €8m to € 15m per
year (RSO/CapGemini estimate from 2009, KPMG estimate from 2010).

− Since obtaining free access in 2011:
o the number of private companies downloading data increased 15-fold;
o alphanumeric data download volume per week increased 20-fold;
o total digital map downloads increased by a factor of 80, (275 to 2,101);
and
o total downloads increased by 100-fold, from 342 to over 3,300.
− ‘Traditional’ re-users have included estate agents, government
departments, businesses commercialising PSI products and services,
organisations and citizens using PSI for non-commercial purposes.
− There has also been an increase in re-use by ‘non-traditional’ entities. One
example is SMEs selling swimming pool products targeting only those
households with a swimming pool.

PSI domain Geographic information

Years of From 2004
measurements

Year of 2011
publication

Sources http://www.etla.fi/files/2696_no_1260.pdf
http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012
/summary.pdf

63

Austrian geographic data case

Austrian Geographic Data

Introduction This case was spotlighted in a Finnish study (Does Marginal Cost Pricing of Public
Sector Information Spur Firm Growth? by Heli Koski, ETLA) that assessed 14,000
firms in the architectural, engineering and related technical consultancy sectors,
located in 15 countries, analysing the effect of maximum marginal cost pricing for
geographical PSI on the firms’ growth performance and the European
Commission’s 2011 POPSIS study (Pricing of Public Sector Information) prepared
by Deloitte and others.

Context − In 2006 the Austrian Federal Office of Metrology and Surveying (Bundesamt
für Eich- und Vermessungswesen (BEV)) made changes to its provision of
geographic data by:
o moving from a pricing structure based on analogue products to one based
on digital products; and
o changing its cost-recovery pricing to marginal cost pricing model.
− The prices of various key data sources, e.g. digital cadastral maps, were
reduced by up to 97%, without additional government funding.
− The pricing regime has been subject to regular reviews and alterations (in
2008 and 2010), based on PSI market value, prices of foreign public sector
bodies for comparable datasets, re-use business conditions, federal
government budget constraints, costs of production and re-use facilitation.

Key findings − The changes are likely to have aided:
o The substantial increase in number of datasets sold:

- In 2007, sales of cartographic products increased by 200%–1,500%,
digital orthophotos by 7,000%, digital cadastral maps by 250%, digital
elevation mode maps by 250%, digital landscape models by 1,000%, and
external-use licences by 100%.
- This increased demand came mostly from Austrian SMEs.
o The 46% increase in total revenue from sales of geographic data.
o The rise in data sought by new user groups, e.g. geo-marketing, location-
based services and health services firms.
o The increase in demand for data from international customers.
− In 2011, purchase orders had stabilised after an initial strong growth period.
But, customer numbers and external licence numbers continue to rise.

PSI domain Geographic information

Years of From 2006
measurements

Year of 2011
publication

Sources http://www.etla.fi/files/2696_no_1260.pdf
http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012/s
ummary.pdf

64

Danish address data case (DECA)

Danish address data

Introduction This case was spotlighted in the ‘Pricing of Public Sector Information Study’ (POPSIS
study) prepared by Deloitte and others for the European Commission. It assessed
models of supply and charging for PSI and implications of price changes.

Context − In 2002 the political decision was taken to establish a central database of all
Danish addresses. This policy change was driven by public task ambitions and
by distinguishing between the public sector investment and subsequent
exploitation of the facility created, allocating the costs to those that benefit,
thus freeing PSBs to rely on cost recovery above re-use facilitation cost level.
− An open network of distributors was established; the network can acquire the
address data against re-use facilitation costs, without any re-use limitations

Key findings − A centrally run system of address data is not only of vital importance for the
proper execution of the public task (such as emergency services, taxation
departments and the monitoring and control of safety regulations), it also
represents an unprecedented source for the private sector to develop and
distribute digital products and services where location is a key element.
− By including the potential returns (in the form of increased economic activities
by the private sector) in the equation when setting up and financing the
database, the maximization of re-use potential (by the private sector) became
a purpose in itself, preventing the PSB from becoming reliant on its own re-use
incomes and allowing it to maximize the multiplier effects downstream.

− The policy change ultimately significantly contributed to:
o a value creation downstream of approximately €57m;
o an increase in FTEs employed by re-users by 800–1,000%;
o an increase in turnover of re-users of around 1,000%;
o a PSB savings of around €5m, against an investment of around €3m;
o an increase in corporate tax gains of around €14m; and
o a return of PSB investment of around 470%.
− The case illustrates that increased tax returns on the boosted turnover of first-
and second-tier re-users downstream in the value chain largely exceed the
investments made by the public sector: the establishment of a central
database of addresses supported by a re-use policy, which only charges
minimal re-use facilitation costs and consequently boosts economic activities
further down the value chain, has financed the more effective performance of
the public task. The PSB has managed to create a self-propelling multiplier
that is available to re-users.

PSI domain Geographic data

Years of 2002–2011
measurements

Year of 2011
publication

Source http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012/su
mmary.pdf

65

Norwegian Meteorological Office case

The Norwegian Met Office

Introduction This case was spotlighted in the ‘Pricing of Public Sector Information Study’
(POPSIS study) prepared by Deloitte and others for the European Commission.
It assessed models of supply and charging for PSI and implications of price
changes.

Context − In 2007 the Norwegian National Meteorological Institute (Met.no) made all
weather data held by it, including certain datasets from other national met
offices, freely available for access and re-use.
o The shift to a zero cost pricing model was mostly financed by
government compensation of €125,000, with the remaining small
amount of transition costs covered by Met.no’s own resources.
o Data is downloaded on a Creative Commons basis.
o Only re-users seeking guaranteed data, for example, full-time
availability, must pay a fee (currently €5,750 per year).
− In 2008 Met.no launched a free online portal (yr.no) which provides graphic
displays of Norwegian and European weather forecasts.
− Met.no still has a commercial arm but this policy is currently under review.

Key findings − The free and open access to this PSI is likely to have aided:
o the average 30-fold increase in the number of re-users, from around
100 to 3,000 unique re-users per week;
o the 200% increase in re-user turnover; and
o the entry of re-users into high end international data provision markets,

normally dominated by national met offices.
− Since 2007 the composition of the re-user mix has changed, with:
o a 40% increase in the number of re-users from outside of Norway; and
o new categories of re-users, including SMEs, that integrate the data into
other media content or use it to develop ‘apps’.
− Increased interaction with re-users/taxpayers have led to:
o improvements in data quality through re-user feedback (Met.no
receives over 5,000 e-mails each year from citizens);
o increased professionalism in service delivery;
o improved internal process efficiency; and
o excellent public relations (Met.no has received awards for being the
most respected public agency for five consecutive years).
− Tax returns on PSI re-use have increased by at least 100% and exceed
revenue loss and the slight uncovered re-use facilitation costs.

PSI domain Meteorological data

Years of 2007–2010
measurements

Year of 2011
publication

Source http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012
/summary.pdf

66

Dutch meteorological data case

Dutch Meteorological Data

Introduction This case was spotlighted in the ‘Pricing of Public Sector Information Study’
(POPSIS study) prepared by Deloitte and others for the European Commission.
It assessed models of supply and charging for PSI and implications of price
changes.

Context − In 1999 the Royal Netherlands Meteorological Institute (het Koninklijk
Meteorlogisch Instituut (KNMI)) decided to cease all its commercial
activities and encourage PSI re-use. Accordingly:
o it approaches PSI provision for re-use as a public task;
o it switched from a full cost-recovery pricing model to recovery of re-use
facilitation costs only; consequently, the re-use system is self-financing;
o the price of the full KNMI dataset has decreased by 80%;
o its commercial arm was reorganised to facilitate privatisation; this was
financed by public funding;
o it does not enrich PSI beyond what is necessary to allow re-use by
public and private re-users alike or produce ‘value-added’ products;
and
o data provision for academic purposes is on a marginal cost basis,
provided that the results of the research are made publically available.

Key findings − Easier access to PSI has led to the creation of a competitive and innovative
private weather market.
− The 1999 policy changes likely aided:

o the 400% increase in turnover for private sector re-users;
o the 250% increase in high-end users;
o a rise in the employment activity of re-users of 300%;
o an increase of over €35m on corporate tax returns;
o the rise of new business models, offering free services to the public
paid through advertising and innovative applications; and
o improvements in data quality, professionalism in service delivery, and
internal process efficiency gains of €3.5m through intensified use and
feedback from re-users.

PSI domain Meteorological data

Years of 1999–2010
measurements

Year of 2011
publication

Source http://ec.europa.eu/information_society/policy/psi/docs/pdfs/report/11_2012
/summary.pdf

67

Australian spatial data case

Australian spatial data

Introduction This case was spotlighted in an Australian study, ‘Costs and Benefits of Data
Provision – Report to the Australian National Data Service’ by John Houghton
from Victoria University, that explores the economic impact of free PSI access
and the costs and benefits. Further information on the case has been drawn
from the 2010 ‘ANZLIC – The Spatial Information Council Economic Assessment
of Spatial Data Pricing and Access’ by PwC.

Context − From 2002, spatial data has been made available as follows:
o online spatial data – free of charge;
o packaged spatial data – marginal cost of transfer; and
o customised spatial data – price not exceeding the cost of transfer.
− No restrictions on commercial use or value-added activities imposed.
− From 2009 data is under a Creative Commons licensing framework.

Key findings − For the year 2009-10, free access to Geoscience Australia’s topographical
(Currency: data was estimated to have increased welfare by $4.7m per annum: data
Australian production costs were around $13.3m with the benefit for government and
dollars) private users being around $10m and $8m per annum, respectively.
− Given the Australian Government’s annual expenditure on fundamental
spatial data of around $70m, free access to such spatial PSI is estimated to
improve welfare by around $25m per annum.
− For the period 2001-02 to 2005-06:
o Free access led to an estimated annual consumer surplus increase of

around $60m.
o There was a good elasticity of demand, with downloads increasing by
up to 172% per annum.
o Average social returns on expenditure were estimated at $15m per
annum.
o Free re-use led to departmental transaction cost savings between 17%
and 33% of respective revenues; conversely, transaction-related cost
savings for PSI re-users were at around $1.7m per annum.

PSI domain Geospatial data

Years of 2001-02 to 2005-06; 2009-2010
measurements

Year of 2011
publication

Sources http://ands.org.au/resource/houghton-cost-benefit-study.pdf
http://www.crcsi.com.au/Documents/ANZLIC-Economic-Study---Stage-2-
Report.aspx

68

Australian statistical data case

Australian statistical data

and the costs and benefits.

Context − From 2005, all publications and data produced by the Australian Bureau of
Statistics (ABS) have been freely available online.
− Between 2005 and 2008, data was made available on a restricted licence
basis; since 2008 all data has been under a Creative Commons (CC) licensing
framework.

Key findings − The cost for the ABS to make statistical data freely available using CC
(Currency: licensing is estimated at around $3.5m per year.
Australian o Savings from staff costs for subscription operations, customer support
dollars) services and processing transaction fees are around $945,000 per year.
o Lost ABS revenue from sales and consultancy are around $4.5m per
year.
o The CC licensing framework has led to estimated internal gains of
$56,667 per year: the framework has low, absorbable, running costs
and revenue forgone has been roughly equal to internal savings.
− However, total user savings are estimated at around $5m per year. This
figure takes account of savings on data and licence costs, transactional
costs, bank costs and time.

− There was a 34% increase in website traffic in the first year of free statistical
data. Product downloads more than trebled between 2003 and 2010.
− Welfare gains have been estimated at around $4m per year.
o This figure is based on the consumer surplus in relation to three-year
average intensity of use for 2003-04, 2005-06 and 2006-07 to 2008-09.
o Due to a lack of information on additional government funding to ABS,
this figure does not account for deadweight loss. However, this is
estimated to be small since the government is a major user of ABS data.

PSI domain Statistical data

Years of 2003–2010
measurements

Year of 2011
publication

Source http://ands.org.au/resource/houghton-cost-benefit-study.pdf

69

Australian hydrological data case

Australian hydrological data

and the costs and benefits.

Context − From 2008 the Australian Bureau of Meteorology (BOM) has published an
annual National Water Account (NWA) that reports on:
o total water resources;
o volume of water available for abstraction;
o rights to abstract water; and
o actual abstraction of water for economic, social, cultural and
environmental purposes.
− Such information was previously difficult to access; in accordance with
regulatory duties it is now provided to BOM by over 200 organisations
holding water data.
− The NWA contains information in standardised forms, is freely available
online and is made available within a Creative Commons (CC) licensing
framework.
o Following BOM encouragement, in 2011 all but nine data providers had
not yet committed to applying CC licensing to the data supplied to
BOM.

Key findings − Water data is used extensively when made readily and freely available
(Currency: online:
Australian o Prior to the introduction of the Victorian Water Resources Data
dollars) Warehouse (VWRDW), an average 400 information requests were made
per year; since 2004 the VWRDW has had an average 443,000 data
requests per year.
o From 2005 to 2010 data extractions for re-use have doubled, from
68,908 to 102,536.
− Open access to information may have been a factor facilitating the
emerging market in water entitlements and allocations that had a value of
$2.8bn in 2009-10.
− Since 2007-08, the equity of the Australian water market has doubled in
value (from around $1.5m to $3m); improved information on water
entitlements, allocations, trade volumes and trade prices is suggested as a
factor in this.

PSI domain Hydrological data

Years of 2005–2011
measurements

Year of 2011
publication

Source http://ands.org.au/resource/houghton-cost-benefit-study.pdf

70

Pettifer 2 study

Pricing of PSI in the Meteorological
Sector blocks market development
Introduction This summary is based on the Position Paper ‘Pricing of PSI in the Meteorological
Sector blocks market development’ by Richard Pettifer, Chairman of the
Association of Private Meteorological Services (PRIMET).

Context − The study notes at the outset that meteorological data is PSI well suited to
being exploited for commercial purposes: it is well known, accessible and
produced in an international format that facilitates quality control, easy
exchange, comprehension and use. Commercial exploitation of
meteorological PSI requires data from a number of states or geographical
regions.
− The study considers three hypothetical SMEs in Luxembourg, Poland and
France that provide weather-related services relating to forecasting,
highways and energy, and uses 2010 prices on a cost-recovery basis.

Key findings − SMEs cannot operate successfully or compete with large firms when partial or
full cost-recovery pricing principles are used: cost-recovery pricing principles
are likely to create barriers to market entry because SMEs are likely to be
unable to find the 20 contracts required to operate profitably.
− The absolute minimum PSI meteorological data required to provide basic
weather-related services, with a market value of €6,000 to €20,000 per
contract, would cost a typical SME between €84,000 and €400,000.
− The loss of revenue from providing meteorological PSI on a marginal cost

rather than a cost-recovery basis is significantly less than the estimated
additional tax revenue on resulting weather-related services.
− If meteorological PSI was provided on a marginal cost basis so that the
market became as well penetrated at that in the US, the estimated gross
overall tax revenue from this weather-related sector would increase by
around €340m.
− Market growth in the weather-related services sector is 15.8% greater in the
US, where meteorological pricing is based on marginal cost principles, than in
Europe, where cost-recovery principles are generally used.

PSI domain Meteorological information

Years of 2002–2010
measurements

Year of 2011
publication

Source http://www.primet.org/file/EU PSI Working Groups/Workshop Position paper
final

71

ANNEX 5 – NATIONAL DATA PORTALS

National data portals

Several countries have set up data portals to give all national users access to country-wide satellite
imagery. These are based on the principle that one government department (often the Ministry of
Agriculture) has been buying data and that, for a relatively small additional cost, it can be made
available to all users through a wider user licence. This arrangement can benefit other government
departments wanting to use the same data as well as industry. Access through a single portal makes
it easier for all users to find the data that they need.

The Dutch National Space Organisation case

Introduction
This is a wrap up of the interview with representatives from the Dutch National Space Organisation
and the Ministry of Economic Affairs, Innovation and Agriculture held on 10 September 2012 in The
Hague.

The arrangement put in place
In March 2012, the Ministry of Economic Affairs, Innovation and Agriculture (the Ministry),
responsible for the Dutch National Space Organisation (the NSO), structurally financed the free re-
use of GMES data for a period of three years (2012, 2013, 2014), by acquiring a licence from Astrium
for radar data covering the Dutch territory. The basic idea is that Dutch users, in particular those
from the private sector, can get acquainted with the use of this dataset, allowing them to prepare for
the GMES data provision in 2014, thus giving them a head start when launched. The components of
the dataset acquired is a look-a-like of the GMES dataset likely to be adopted (although it currently

also holds two meter pictures). The total funding amounts to €4m for the entire period and was
furnished by the Ministry, by rearranging funds within its own budget (rather than by claiming extra
money from the Dutch Treasury, which would have been very burdensome in these times).

Only users based in the Netherlands are allowed to re-use the data. To that end they have to register
with NSO and there is a ‘light’ check on the registration details. The data processing and making
available is all done by Astrium, NSO stays clear of that. Also the user enters into a licensing
agreement with Astrium, not with NSO. So far, around 250 accounts have been created (probably
with a multitude of users per account). The amount is steadily growing (currently about five new
accounts per week). Interestingly, Astrium also chipped in the two meter resolutions for free.

How did NSO manage?
The movement was created bottom-up – the NSO had been contemplating this possibility for a long
time. To that end it had already initiated a study, looking at how the private sector was making its
money and assessing the effects that such movement would have. Interestingly, the study found that
at that time, the private sector was relying heavily on subsidized income streams, rather than on
income generated from turnover. Obviously, setting up, allocating and monitoring these subsidy
streams created a lot of overhead both for the government and the private sector. Further, the costs
to acquire the data were (and are) quite substantial, creating a significant entry barrier for the
private sector, SMEs in particular. However, a subsidized space program had been established in the
1980s, creating a fairly robust knowledge infrastructure. And of course, the Dutch Government had
been investing in this field for many years through ESA contributions. (The total amount of Dutch
governmental cash out is around €100m per year). Put differently, investments had been made, the

72

basic infrastructure was there, the market opportunities are there, but there was a gap between
knowledge and market exploitation thereof. This created the business case.

It all came together when the former Ministry of Economic Affairs and that of Agriculture and Fishery
merged into a new Ministry (the Ministry of Economic Affairs, Innovation and Agriculture). Suddenly
the public sector and private sector user interests were bundled. The NSO, together with supporters
within the Ministry, managed to set up the business case and pitch the idea appropriately, also
connecting it to the ‘Top Sectors’ policy adopted by the Ministry, which defined a number of focal
areas of strategic importance for the Dutch economy. Thus, the NSO managed to convince the
decision makers that continuity is key, particularly for the market parties, thus validating a claim for
structural funding for a period of three years. Setting up the business case, the NSO also mobilized
20–25 private sector users, requesting them to match the investment. Ultimately, private sector
users committed €20m as own investments, to be spent over time once the data was made available.
This was an important winning argument towards the decision makers. The NSO’s case was further
strengthened by aligning leading researchers who backed the business case. So clearly this was a
bottom-up movement that would not have succeeded without the continuous efforts to make it
work, in combination with smart tactics and a bit of luck.

Further actions needed
What is needed now is the visibility of the services that are emerging, based on the data. To that end
the NSO is also pushing the usage by stimulating public sector users publicly procuring services that
could be based on GMES to put that requirement in the tender specifications, thus turning the public
sector into a launching customer. Furthermore, the NSO expects that the combination with in situ
data will result in many new services.

The Swedish data portal case

The Swedish data portal was launched in June 2008 after a parliamentary decision in 2007. Five
government agencies (the Environmental Protection Agency, the National Land Survey, the Swedish
Meteorological and Hydrological Institute, the Swedish Forest Agency and the Swedish National
Space Board) made a combined initiative to suggest the establishment of a satellite image database,
and in 2007 a specific budget was granted for its development and for data procurement. The
government also granted a specific budget for data procurement in 2008.

The portal is run by Lantmäteriet – the Swedish mapping, cadastral and land registration authority –
with Metria providing the technical support. Lantmäteriet's running costs are normally covered by
charging user fees. However, for the EO Portal, a consortium currently comprising the Environmental
Protection Agency, the Swedish National Space Board, the Swedish Forest Agency, the Swedish
University of Agricultural Sciences and four forest companies is meeting the costs.

The database contains satellite imagery for different types of users. It includes historical datasets
from 1970, 1980, 2000 and 2005 of complete country coverage at between 10m and 30m. System
corrected and ortho-rectified imagery are available for professional users while the general public
might be more interested in image mosaics to see how the landscape has changed over time.

From 2007, full country coverage will be made each year. The goal is to make these datasets easily
accessible by anyone. By early 2009, there were over 1250 unique users with 10 new users
registering each week. In 2010, this had reached 2000 registered users, and 2500 in 2011.

The history of downloads shows a sharp increase in data usage. In 2008-09 there were 6500 scenes
taken, as compared to around 300 scenes in the years before the portal was established. In 2010, this

73

increased to 21,000 scenes and there was a further small increase in 2011 to an estimated 25,000
scenes. Around 80% of the products being downloaded are ortho-corrected.

Foretag och NGO’s = company and NGO’s In blue = total number of users (left)
Mindigheter = agencies In red = total downloads (right)

A re-user perspective: Metria (Sweden)

Metria is one of the leading EO service companies in Sweden, having benefited from the free data
access policy. The company situation is rather unique: it was an operating entity of the Lantmäteriet
until 2011 when it became a separate company. Prior to the Portal, Metria was the only organisation
that had access to the imagery bought by the Swedish public sector. This placed it in a privileged
position as it could use data from these archives for other projects. Once the Portal was established
and anyone had access to the imagery, Metria found that this helped them win business through
increased transparency. Customers (and competitors) now knew the transfer pricing policy, which

previously was not visible.

Overall, service providers in Sweden are estimated to have gained annual revenues of more than
€1m (with Metria gaining around €700,000) as a result of the Portal and the free access policy. The
main uses of the data have been in forestry, environmental monitoring and urban planning.

Norwegian Case

Norway has also implemented a centralised approach where a bulk data purchase from the Canadian
radarsat has complemented data coming from Envisat before it died. These data were made
available for government departments to use, mostly for coastal patrols, ice and oil spills.
Commercial customers are required to purchase any data needed for their purposes. Most recently,
the Norwegian receiving station has been active in supplying oil spill detections to the European
Marine Safety Agency as an initial operation under the foreseen GMES services.

74

ANNEX 6 – LIST OF CONTRIBUTORS

We kindly thank all those that have contributed to the study, in particular:

Representatives from ESA

Josef Aschbacher and Alessandra Tassa

Workshop participants

EARSC Workshop (6th July 2012) Expert Workshop (7th November 2012)

Name Organisation Country Name Organisation
Miles Taylor Aerodata International Belgium Reinhard Schulte Braucks EC-DG Enterprise
Jon Styles Assimila UK Josef Aschbacher ESA
Thomas Schrage Astrium Services Germany Francesco Barbosa EC-DG Connect
Paul Stephens DMCII UK Vark Helfritz Astrium Services
Han Wensink BMT-Argoss Netherlands Marcello Maranesi E-geos
Antonio de Sousa Edisoft Portugal Frederica Mastracci E-geos
Fabrizio Pirondini Elecnor Deimos Spain Alan Belward EC-JRC
Frederic Collomb FDC France Par Astrand EC-JRC
Thomas Kukuk GAF Germany Ian Davison EEA
Christian Hoffman Geoville Austria Catharina Bamps Eurogi
Vital Schreurs GIM Belgium Guro Dahle Strom Norwegian Space Office
Celestino Gomez GMV Spain Joern Hoffmann DLR
Joerg Schaeffer IABG Germany Alessandra Tassa ESA
Emilio Viedma Indra Spain Stephen Coulson ESA
Chetan Pradhan Logica UK Rudy Aernoudt EC DG Enterprise

Erik Willen Metria Sweden Deborah Sciocco EC DG Enterprise
Massimiliano Vitale RapidEye Germany Jansky Radomir EC DG Enterprise
Sylvie Remondiere Serco Italy Herve Jeanjean EC DG Enterprise
Claire-Anne Reix Thales Alenia Space France Olivier Chassagne EC DG Enterprise
Hugo de Groof DG Environment Daniel Quintart EC DG Enterprise
Jerome Bequignon ESA Torsten Riedlinger EC DG Enterprise
Stephen Coulson ESA Francoise Villette EC DG Enterprise
Denis Bruckart EUSC Ingo Baumann BHO Partners
Ian Davison EEA Olivier Heinrich BHO Partners
Richard Swetenham EC DG Connect Stephane Ourevitch Spacetec Partners
Richard Pettifer PRIMET Valerie Forlin Spacetec Partners
Jossiane Masson EC-DG Enterprise Dimitrios Papadakis Spacetec Partners
Marc de Vries Citadel Consulting Carla Filotico Spacetec Partners
Geoff Sawyer EARSC Marc de Vries Citadel Consulting
Monica Miguel-Lago EARSC Geoff Sawyer EARSC

75

ANNEX 7 – BIBLIOGRAPHY
8. Berenschot/Netherlands Economic
Institute (2001), Welfare implications of
The books, articles and reports listed below different pricing
are the main sources on the economic aspects models for public sector information
surrounding use and re-use of public sector (Eindrapport Welvaartseffecten van
information. While drafting this report, a verschillende
substantial amount has been consulted. financieringsmethoden van elektronische
gegevensbestanden), Berenschot and NEI,
1. Aichholzer, G. and P. Tang (2004), Utrecht: Berenschot.
Harnessing public sector information for
general accessibility: Austria and the UK, 9. BETA University of Strasbourg (2011), The
in Aichholzer, G. and H. Burkert (Ed.), reuse of PSI – An economic optimal pricing
Public Sector Information in the Digital model, Agence du patrimoine immatériel
Age: Between Markets, Public de l'État.
Management and Citizen’s Rights,
Cheltenham UK: Edward Elgar. 10. Bing, J. (1998), Commercialisation of
Geographic Information in Europe,
2. Alani, H. (Ed.) (2007), Unlocking the Norwegian Centre for Computers and Law.
potential of public sector information with
semantic web technology, Lecture Notes 11. Bird, R. M. and T. Tsiopoulos (1997), User
in Computer Science, 4825, 708. Charging in the Federal Government - A
Background Document, Ottawa: Treasury
3. Allan, R. (2009), The Power of Government Board.
Information. In J. Gøtze and C. B.
Pedersen (Ed.), State of the eUnion: 12. Blomquist, S. and V. Christiansen (2005),
Government 2.0 and Onwards, The role of prices for excludable public

AuthorHouse. goods, International Tax and Public
Finance 12 (I).
4. Amos J. W. (1999), Freedom of
Information and Business, Appendix on 13. Bundesministerium für Wirtschaft und
USA and FOIA Agency Technologie (2008), Chancen für
case study, The Constitution Unit, Geschäftsmodelle deutscher
University College London. Unternehmen im europäischen und
globalen Geoinformationsmarkt, MICUS
5. Arrow, K. J. (1984), Information and Management Consulting GmbH.
economic behaviour, Collected Papers of
K. J. Arrow, Vol. 4, The Economics of 14. Cabinet Office (1998), Crown Copyright in
Information, Cambridge MA: Harvard the Information Age (Green Paper),
University Press. London. Stationery Office, United
Kingdom.
6. Arrow, K. J. (1979), The economics of
information, in Dertouzos, M.L. and J. 15. Centre of Land Policy and Valuations of
Moses, The Computer Age: A Twenty Year the Universitat Region of Catalunya
View, Cambridge MA: MIT Press. Politècnica de Catalunya (2007), Study of
the Economic Impact of the Spatial Data
7. Barr, R. and I. Masser (1996), The Infrastructure.
Economic Nature of Geographic
Information: A Theoretical 16. Corbin, C. (2003), New Issues for the
Perspective, Proceedings of the GIS European GI Strategy: Public Sector
Research UK Conference. Information, Sheffield: University of

76

Sheffield, GINIE (Geographic Information Sciences Geographica Polonica, Vol 80, No
Network in Europe). 2.

17. Corbin, C. (2007), Public Sector 25. Dutch Ministry of the Interior and
Information – Financial impact of the PSI Kingdom relations (2000), Towards
Directive: Pricing and Charging. optimum availability of public sector
information, The Hague.
18. Corbin C. (2010), Public Sector
Information: Economic Indicators & 26. European Commission (2003), Directive
Economic case study on charging models, 2003/98/EC, European Parliament and
ePSIplatform. Council Directive of 17 November on the
re-use of public sector information.
19. Craglia, M. and K. Evmorfopoulou (1999),
Geographical data infrastructures four 27. Evans, P. and T. Wurster (2000), Blown to
European case studies, Department of bits: How the new economics of
Town and Regional Planning, University of information transforms strategy, Harvard
Sheffield. Business School Press.

20. De Vries, M. (1999), Initiatives in EU 28. Hadi, Z. A., and N. McBride (2000), The
Member States in respect of Commercialisation of public sector
dissemination and exploitation of Public information within UK government
Sector Information. departments, International Journal of
Public Management 13 (7).
21. De Vries, M., Van de Velde, R., Dekkers,
M., and Polman, F. (2006), Measuring 29. Hall, M. (2003), Spatial Data
European Public Sector Information Infrastructures in Australia, Canada and
Resources, on behalf of DG Information the United States, Leuven: K.U. Leuven &

Society, European Commission. M. Hall consultant. Report elaborated in
the context of a study commissioned by
22. De Vries, M., Kappf, L., O’Connor, J., the EC (EUROSTAT & DGENV) in the
Negreiro Achiaga, M., Wauters, P., Osimo, framework of the INSPIRE initiative.
D. (2011) Pricing of Public Sector
Information Study, on behalf of DG 30. H. M. Treasury (2000), The Economics of
Information Society, European Government Information, in Cross-Cutting
Commission. Review of the Knowledge Economy, Part 5
London: HM Treasury, United Kingdom.
23. De Vries, M., Lundqvist, B. (2011) Missing
the Link? Interpreting the Terms ‘Business 31. H. M. Treasury (2002), Selling Government
Activities’ and ‘Exclusive Rights’ in the Services into Wider Markets: A Policy Note
Swedish Act on Re-use of Public Sector for Public Bodies, London: HM Treasury,
Information, Against the Background of United Kingdom.
the EU Directive on Re-use of Public
Sector Information, on behalf of the 32. Hookham, C. (1994), The Need for Public
Swedish Competition Authority Sector Policies for Information Availability
& Pricing, Cambridge Computer
24. Donker, T. W. (2007), Access to and Re- Consultants (UK) Ltd.
use of Public-sector Environmental Data
and Information. Policy Developments 33. Katz, R. W. and Murphy, A.H. (1997),
with a Focus on the European Hydro- Economic Value of Weather and Climate
Meteorological Scene, Polish Academy of Forecasts, Cambridge, Cambridge
University Press.

77

National Academies, National Academy
34. Lazo, J. K. and L. G. Chestnut (2002), Press.
Economic Value of Current and Improved
44. Newbery, D. (Ed.) (2008), Models of Public
35. Weather Forecasts in the US Household Sector Information Provision via Trading
Sector, In Report prepared for the Funds, Cambridge University.
National Oceanic and Atmospheric
Administration by Stratus Consulting Inc. 45. Nilsen, K. (1997), Social Science Research
Boulder, Colorado. in Canada and Federal Government
Information Policy: the Case of Statistics
36. Lind, M. (2003), Reliable Address Data: Canada, Dissertation. University of
Developing a Common Address Reference Toronto.
System, National Survey and Cadastre
Denmark. 46. Nilsen, K. (2001), the Impact of
Information Policy: Measuring the Effects
37. Lind, M. (2008), Addresses as an of the Commercialization of Canadian
infrastructure component: Danish Government Statistics, Westport CT:
experiences. National Survey and Cadastre Ablex.
Denmark.
47. Nilsen, K. (2007), Economic Theory as it
38. Lind, M. (2010), The value of Danish Applies To Statistics Canada: A Review of
address data: Social benefits from the the Literature, PhD, University of Toronto.
2002 agreement on procuring address
data etc. free of charge,” Danish 48. Office of Fair Trading (OFT) (2006), The
Enterprise and Construction Authority. Commercial Use of Public Information
(CUPI), DotEcon Ltd.
39. Loenen, B. (2006), Developing geographic

information infrastructures. The role of 49. Office of Fair Trading (OFT) (2006), The
information policies, Delft, DUP Science. Commercial Use of Public Information,
Annexe G: Economic value and detriment
40. Longhorn, R. and M. Blakemore (2004), analysis.
Re-visiting the Valuing and Pricing of
Digital Geographic Information, Journal of 50. Pas, J. and B. Vuyst, de (2004), Re-
Digital Information, 4(2). establishing the balance between the
public and private sector: regulating public
41. Mayo, E. and T. Steinberg (2007), The sector information commercialization in
Power of Information: an independent Europe, Journal of information Law and
review. Technology.

42. MICUS Management Consulting GmbH 51. Pettifer, R. E. W. (2008), Towards a
(2008), Assessment of the Re-use of Public Stronger European Market in Applied
Sector Information (PSI) in the Meteorology, Meteorological Appl., Vol
Geographical Information, Meteorological 15.
Information and Legal Information
Sectors. 52. Pira International Ltd., University of East
Anglia, and KnowledgeView Ltd. (2000),
43. National Research Council (2003), Fair Commercial Exploitation of Europe's
Weather: Effective Partnerships in Public Sector Information, European
Weather and Climate Services, Committee Commission, Directorate General for the
on Partnerships in Weather and Climate Information Society.
Services, National Research Council of the

78

53. Pollock, R., D. Newbery and L. Bently
(2008), Models of Public Sector 63. U.S. Departments of Commerce and
Information Provision via Trading Funds, Transportation (2006), Beneﬁts of the
BERR (commissioned by HM Treasury and New GPS Civil Signal: The L2C Study,
BERR). Leveson Consulting.

54. Ravi Bedrĳvenplatform (2000), Economic 64. U.S. Department of Labour (2004),
effects of low entry accessibility to public Geospatial Technology high growth
sector information (Economische eff ecten industry profile, Employment and Training
van laagdrempelige beschikbaarstelling Administration, Business Relation Group.
van overheidsinformatie), Amersfoort:
Ravi. 65. Vickery, G. (2011), Review of recent
studies on PSI re-use and related market
55. Richard M. Adams, R. M., L. L. L. Houston, developments, Report for the European
R. F. Weiher (2004), The Value of Snow Commission, 41 pp.
and Snow Information Services, Report
prepared For NOAA’s National 66. f, P. and P. Backlund (1997), International
Operational Hydrological Remote Sensing Information Policy and Conflict: Open and
Center. Unrestricted Access versus Government
Commercialization, in Kahin, B. and C.
56. Robinson, D. (Ed.) (2008), Government Nesson (Ed.), Borders in Cyberspace:
data and the invisible hand, Yale Journal Information Policy and the Global
of Law and Technology, 11, 160. Information Infrastructure, Harvard
Information Infrastructure Project,
57. Saulles, de, M. (2005), e-Government and Cambridge MA: MIT Press.
the Re-use of Public Sector Information in
5th European conference on e- 67. Weiss, P. (2002), Borders in Cyberspace:

government: University of Antwerp, Conflicting Public Sector Information
Belgium. Policies and Their Economic Impact,
Summary Report, Washington: National
58. Stiglitz, J. E. (1999), Knowledge as a Global Oceanic and Atmospheric
Public Good, Washington: World Bank. Administration/National Weather Service,
Department of Commerce.
59. Stiglitz, J. E. (2000), Economics of the
Public Sector, 3rd ed. New York: W. W. 68. Weiss, P. (2003), Conflicting International
Norton. Public Sector Information Policies and
Their Effects on the Public Domain and the
60. Stratus Consulting Inc. (2002), Economic Economy, in Esanu, J. M. and P. F. Uhlir,
value of current and improved weather Role of Scientific and Technical Data and
forecasts in the U.S, SC10050. Information in the Public Domain,
Proceedings of a Symposium, Board on
61. Uhlir, P. (2009), The Socioeconomic International Scientific Organizations,
Effects of Public Sector Information on Washington: National Academies Press.
Digital Networks: Toward a Better
Understanding of Different Access and 69. Weiss, P. (2004), Borders in cyberspace:
Reuse Policies: Workshop Summary. Conflicting government information
policies and their economic impacts No.
62. UNESCO (2004), Policy guidelines for the 17, in Esanu, J. M. and P. F. Uhlir, Open
development and promotion of Access and the Public Domain in Digital
governmental public domain information, Data and Information for Science,
Uhlir, P. (CI-2004/WS). Proceedings of an International

79

Symposium, Board on International
Scientific Organizations, Washington:
National Academies Press.

70. Zevenbergen, J. (Ed.) (1998), Free
Accessibility of Geo-information in the
Netherlands, the United States and the
European Community: proceedings of the
seminar held on October 2, 1998 at Delft,
the Netherlands.


80

EARSC Open data study - Final Report

More Related Content

Similar to EARSC Open data study - Final Report

More from Planetek Italia Srl

Recently uploaded

EARSC Open data study - Final Report