ISSUE 2008/03

FEBRUARY 2008
bruegelpolicybrief

EUROPE’S R&D: MISSING

THE WRONG TARGETS?
by Bruno van Pottelsberghe SUMMARY Europe is not delivering on its Lisbon agenda commitment to
Senior Fellow at Bruegel
Professor at ULB-ECARES
increase its R&D-to-GDP ratio to three percent by 2010. This is worrying,
bvp@bruegel.org not only because Europe seems unable to reach an objective it has publicly
set itself, but mainly because in 2006 its R&D intensity was still below two
percent, having flatlined for more than two decades. As far as business-
funded R&D is concerned, the Chinese business sector has even
outperformed European firms. The Lisbon-inspired national R&D targets are
equally overambitious. The European Commission’s benchmarking of mem-
ber states against the headline three percent figure is questionable
because such comparisons rarely take into account the effect of industrial
specialisation. For most countries, R&D intensity is a by-product of special-
isation. However, Swedish and US R&D intensity is higher than their
industrial structure would suggest, implying that other factors are at work,
such as a large integrated technology market and a superior academic
Business-funded R&D as % of
research environment.
GDP, 2000 and 2006 (or closest) POLICY CHALLENGE

At EU level, the aggregate government

China 2006 sector should first correct its own
2000
failure and support research activities
up to a threshold of one percent of GDP.
EU27* Setting targets for private R&D is inef-
fective. The drivers of private R&D call
United
for a more integrated European market
States for technology, notably an EU patent in
lieu of the current system, which
involves prohibitive costs. Also, more
Japan*
funding is needed for academic
0.0 0.5 1.0 1.5 2.0 2.5 3.0
research, as a magnet for local and for-
eign business R&D activity in Europe.
Source: OECD MSTI, 2007. Industry-financed GERD as a %
of GDP; * indicates the year 2005 instead of 2006.
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

ONE OF THE MAIN GOALS of the EU’s

02 Lisbon agenda is to achieve a high-
er level of research and 3.50
Figure 1: R&D intensity of selected EU and non-EU countries, 2006
bruegelpolicybrief

development (R&D) spending. Two

3.00
sub-targets for R&D spending were
clearly defined in 2002: EU R&D
2.50
intensity (R&D expenditure divid-
ed by GDP) was to increase from
2.00
about 1.8 percent in the late
1990s to about three percent by China
1.50
2010; and two-thirds of this EU-15
spending was to be funded by the EU-27
1.00 Japan
business sector, the rest being
United States
funded by governments. 0.50 1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
As illustrated in Figure 1, R&D
intensity in the EU has been stable Source: OECD, MSTI, 2007. The figures are gross expenditures on R&D as a percentage of GDP. The
2006 figures for the EU have been extrapolated from Eurostat figures. OECD sources are used
since the early 1980s, fluctuating because they provide comparable figures for China, Japan and the US.
between 1.6 percent and 1.8
percent. In 2006, R&D intensity in counter-intuitive, behaviour with percent threshold. Sweden’s per-
the EU was still under 1.8 percent. regard to their own self-set agen- formance lies close to four
The relative spend on research da. In section two we explain why percent. Denmark, Austria and
activities in the US has also been common R&D targets make little Germany are around the 2.5
stable, but on average above 2.5 economic sense in an EU where percent threshold, whereas France
percent. Japan exhibits an impres- industrial specialisa- is just above two
sive performance, with a constant- tion differs substan- percent. However, the
ly increasing R&D intensity that tially across countries. vast majority of
has remained well above three Failing to account for ‘Europe’s R&D effort countries has an R&D
percent since the early 2000s. national industrial has been flatlining intensity of well below
Figure 1 also illustrates the dra- structures may for two decades.’ two percent, fluctuat-
matic increase in China’s total actually lead to ing between 0.5
R&D expenditure relative to GDP, skewed country percent and two
from about 0.5 percent 10 years benchmarks. Section percent of GDP, with a
ago to 1.5 percent in 2006. Bottom three investigates what can be median of 1.2 percent. This broad
line: the EU is not really catching done to improve the expected range of intensities is also
up with the US or Japan in terms of return to R&D in Europe, and observed within the US, but with a
research spend, while China is hence the propensity to invest in median R&D intensity that is
catching up with the EU. R&D. It sets out two broad policy much higher than in Europe, as
recommendations which would illustrated in Table 1. The best
The objective of this Policy Brief is improve Europe’s R&D prospects. European performer, Sweden, has
to provide a critical assessment of an R&D intensity which is less
the R&D component of the Lisbon 1. DELIVERY FAILURE than half that of the top US per-
1
New Mexico is a agenda. Section one underlines former, New Mexico1. Seven US
relatively small state,
which has a remarkably the considerable gap between the The intensity of R&D spending states have an R&D intensity high-
high level of R&D current levels of R&D intensity and across EU member states varies er than four percent, against none
intensity. This is largely
attributable to federal
the national objectives that were considerably. Figure 2 shows that for the EU.
support to federally announced as part of the relaunch some countries have reached
funded R&D centres of the Lisbon agenda. This section relatively high levels, especially Trends in the R&D-to-GDP ratio pro-
(FFRDCs) provided by
the US Department of also illustrates governments’ Finland and Sweden, which several vide an interesting insight into
Energy. sluggish, and in certain cases years ago leapfrogged the three how active countries have been in
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

Table 1 actually devotes one percent of

Structure of R&D intensity across the EU and US states, most recent data
EU27, 2006 US, 2004
its GDP to funding public (high-
er education, laboratories) or
03

bruegelpolicybrief
business-channelled (subsi-
Maximum Sweden 3.8% New Mexico 8.0%
dies and procurement)
Cyprus (0.42%) Wyoming (0.40%) research activities. The only
Minimum
Romania (0.46%) South Dakota (0.50%)
countries that are close to the
Median across states 1.2% 1.9% one percent target are Sweden,
Average across states 1.4% 2.2% Austria and Finland.
90th percentile 2.5% 4.3% • Second, despite the Lisbon
Source: Bruegel based on Eurostat, US National Science Foundation, Division of Science Resources agenda, a large number of
Statistics, National Patterns of R&D Resources (annual series), Science and Engineering Indicators countries have actually
2007. The full state-level data is presented in Figure 2.

seeking to improve their relative Fig 2. R&D intensity of US federated states (2004), and EU member states (2006)
performance. From 1996 to 2006 Wyoming
Cyprus
the median R&D intensity in Romania
Bulgaria
Europe increased only slightly. In Slovakia
South Dakota
absolute terms, the most dynamic Malta
Poland
Greece
countries have been Finland (+1.2 Louisiana
Arkansas
percent), Austria (+0.9 percent), Nevada
Latvia
Denmark (+0.6 percent) and Oklahoma
Alaska
Sweden (+0.5 percent). These four Kentucky
Lithuania
Portugal
countries already had a very high Mississippi
Maine
level of R&D intensity in 1995, and Florida
Hawaii
have made the most marked Hungary
West Virginia
improvement over the subsequent Montana
Georgia
Nebraska
decade. It is worth mentioning that Italy
Estonia
three countries have seen a drop Spain
South Carolina
in their levels of R&D intensity: Ireland
New York
Tennessee
France (-0.1 percent), the United Iowa
Missouri
Kingdom (-0.2 percent), and the Czech Republic
Luxembourg
Netherlands (-0.2 percent). Their Texas
Slovenia
levels of R&D expenditure play an Netherlands
United Kingdom
Wisconsin
important part in aggregate EU Arizona
Belgium
R&D intensity. As illustrated in Ohio
Utah
Table 2 (overleaf), this drop may Alabama
North Carolina
be explained in part by a drop in Illinois
France
Kansas
government-funded R&D observed Indiana
Virginia
in the three countries, which has Delaware
Idaho
not been offset by an increase in Pennsylvania
Denmark
Austria
business R&D expenditure. Table 2 North Dakota
Vermont
presents government-financed Germany
Minnesota
and industry-financed R&D as a Oregon
Colorado
percentage of GDP, in 1995 and New Jersey
District of Columbia
New Hampshire
2006. Three main observations Finland
Sweden
can be made about these figures: California
Washington
Connecticut
Rhode Island
Michigan
• First, none of the EU member Massachusetts
Maryland
states has fulfilled its self-set New Mexico

commitment, as no country 0 1 2 3 4 5 6 7 8

Source: Bruegel based on Eurostat, National Science Foundation, Division of Science Resources
Statistics, National Patterns of R&D Resources (annual series), Science and Engineering Indicators 2007
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

reduced their government fund- witness to the dramatic further away from the Lisbon tar-
04 ing of R&D as a percentage of
GDP. The aggregate EU27
increase in private R&D activity
in China.
get a country was, the bigger the
increase projected in the national
bruegelpolicybrief

government-funded R&D programme implementing the

intensity fell between the mid In addition to this counter-intuitive Lisbon agenda. Although this could
1990s and 2005. Interestingly, behaviour whereby the ‘average’ be seen as expressing political will
a drop also occurred in the US EU government has actually to catch up with the best perform-
and Japan over the same reduced its support to R&D activity ers, many of the targets set are
period, but it was largely com- over the past ten years, the spend- clearly unrealistic. They appear to
pensated for by a more than ing targets which individual represent wishful thinking rather
proportional increase in busi- countries have chosen to set than political momentum. The
ness-funded R&D, which was themselves were overly ambitious. right-hand side of Figure 3 shows
not the case for EU27. Indeed, Figure 3 illustrates a clear that some countries have set
• Third, the Chinese business- positive relationship between a 2010 targets that are between two
funded R&D intensity is at the country’s distance from the three and four times higher than their
same level, in fact a little higher, percent target in 2004 and the tar- level of R&D intensity in 2004.
than that of Europe, bearing get it has set itself for 2010. The
2. SKEWED COUNTRY
BENCHMARKING
Table 2
Industry and government-financed gross expenditure on R&D (GERD), as
a percentage of GDP (1995 and 2006, or closest date) In addition to the relative
government spend on research
Industry-funded GERD Government-funded GERD
activities, a second issue that
2006 1995 Difference 2006 1995 Difference
must be examined when
Sweden 2.55 2.17 0.38 0.91 0.96 -0.05
evaluating countries’ R&D per-
Finland 2.30 1.35 0.95 0.87 0.79 0.08
formance is industrial specialisa-
Germany 1.68 1.31 0.37 0.70 0.83 -0.13
tion. A country specialised in
Denmark 1.46 0.82 0.64 0.67 0.72 -0.05
finance (eg Luxembourg) would
Luxembourg 1.28 na na 0.27 na na
not need a high level of R&D
Austria 1.14 0.70 0.44 0.90 0.72 0.18
expenditure in order to ensure
France 1.12 1.10 0.02 0.82 0.96 -0.14 growth – at least as commonly
Belgium 1.11 1.12 -0.01 0.46 0.39 0.07 measured (the innovative efforts
Netherlands 0.90 0.90 0.00 0.64 0.83 -0.19 that are required to introduce new
Czech Republic 0.88 0.60 0.28 0.60 0.31 0.29 financial products are not included
Slovenia 0.82 0.72 0.10 0.56 0.64 -0.08 in R&D statistics). Similarly, a
Ireland 0.79 0.85 -0.06 0.40 0.28 0.12 country specialised in tourism,
United Kingdom 0.75 0.94 -0.19 0.58 0.64 -0.06 fashion, services or food would
Spain 0.52 0.35 0.17 0.48 0.35 0.13 logically have lower R&D intensity
Italy 0.43 0.41 0.02 0.56 0.52 0.04 than a country specialised in the
Hungary 0.43 0.27 0.16 0.45 0.38 0.07 pharmaceuticals, engineering or
Portugal 0.29 0.11 0.18 0.45 0.35 0.10 biotech industries. Interpretations
Poland 0.18 0.23 -0.05 0.32 0.38 -0.06 drawn from Figure 2 and Figure 3
Slovak Republic 0.17 0.55 -0.38 0.27 0.35 -0.08 are therefore to be treated with a
Greece 0.16 0.10 0.06 0.24 0.20 0.04 substantial degree of caution. For
Romania 0.14 0.31 -0.17 0.29 0.46 -0.17 instance, Finland has a reputation
Median 0.82 0.70 0.10 0.56 0.46 0.00 for specialisation in information
EU27 0.94 0.86 0.08 0.61 0.66 -0.05 and communication technologies,
United States 1.70 1.51 0.19 0.77 0.89 -0.12 an industry which is very intensive
Japan 2.53 1.96 0.57 0.56 0.67 -0.11 in R&D. Taking into account this
China 0.99 na na 0.35 na na specialisation, the Finnish R&D
Source: OECD, MSTI, 2007.
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

In order to evaluate the extent to

Figure 3: R&D intensity targets for 2010 compared ‘the Lisbon gap’ in 2004

Ratio of the national

4
LV
which industrial specialisation
may affect our assessment of
05

bruegelpolicybrief
target (2010) to the PL
national R&D national R&D performance we rely
intensity (2004)
3
on the estimates provided by
LT
SK
Mathieu and van Pottelsberghe
HU CY
IT EE (2008), which seek to shed light
PT
2 SL
GR on the drivers of business-funded
LU ES
NL
R&D at the industry level. They use
IE
DK BE
FR UK panel data of industry-specific
FI DE AT R&D spending for about 20
1
industrial sectors in 10 countries
over the period 1991-2002. Their
0 results lead to three observations:
-1 -0.5 0 0.5 1 1.5 2 2.5 3
Gap between the 3 percent target and the level of R&D intensity in 2004
• Technological specialisation
Source: Bruegel based on European Commission, National Reform Programmes and annual reports explains the variation in R&D
on implementation. Bulgaria, the Czech Republic, Romania and Sweden not included as we were not
able to find an explicit R&D target in the National Reform Programmes of those countries.
intensity much better than any
other country specificities.
intensity may be perceived as not R&D intensity may actually not be • Not taking into account
being particularly high. performing particularly well given industrial specialisation may
their specialisation in R&D-inten- lead to a highly skewed ranking
The role of specialisation has sive industries. Figure 4 shows the of countries.
received increased attention in R&D intensity of most • When industrial specialisation
recent European reports on manufacturing industries aver- is taken into account, only
innovation (see the Aho Group aged over ten OECD countries. It is Sweden and the US still outper-
report (2006), the second report clear that there are very consider- form other countries. Neither
of the Knowledge for Growth Group able differences between sectors. Japan nor Finland has a partic-
(2007) and the Commission’s Key This confirms that international ularly high R&D intensity in
Figures 2007). This is important, comparisons of R&D intensities relation to what their industrial
as some countries generally should take account of the particu- structures would suggest.
praised for their above-average lar specialisation of each country.
In a nutshell, business R&D
Figure 4: R&D intensity by industry, average across ten countries intensity is endogenous, not
Services sector
exogenous. Governments should
Paper, paper prod. & printing therefore go beyond traditional
Metal products
Food, beverages & tobacco
incentive policies such as direct
Textiles, apparel & leather R&D subsidies or tax credits. To set
Petroleum refineries & product
Non-metallic mineral products
a business-funded R&D target at
Wood products, furniture, other manufac., nec the country level is thus either
Iron & steel
Non-ferrous metals
wishful thinking or an implicit
Shipbuilding & repairing industrial policy – a way to alter
Rubber & plastic products
the country’s industrial structure.
Chemicals excl. drugs
Other transport equipment In other words, there is no basis for
Non-electrical machinery
the setting of EU-wide or country
Motor vehicles
Aircraft targets in the Lisbon programmes
Professional goods
unless the EU’s intention is to
Drugs & medicines
Office & computing machinery determine member states’
Sub-total electrical-electronical industrial structure. Pouring R&D
0% 5% 10% 15% 20% 25% 30% money into low-tech sectors would
Source: Mathieu and van Pottelsberghe based on OECD, ANBERD and STAN databases (2005).
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

clearly have only a very small return on the investment. What and enforced in all relevant nation-
06 impact on aggregate efficiency. would improve this expected
return? Beside fashionable R&D
al patent offices. For that reason, a
patent examined by the EPO and
bruegelpolicybrief

The strong increase observed in tax credit or direct subsidisation then enforced in 13 European
the R&D intensity of Finland, policies designed to reduce the countries costs about 11 times
Denmark and Sweden is attributa- cost of carrying out R&D, two spe- more than a patent granted by the
ble in large measure to the trend in cific policy areas deserve particu- United States Patent and
their technological specialisation lar attention in Europe4: Trademark Office (USPTO), and 14
towards R&D intensive industries, times more than a patent granted
as illustrated in European An integrated market for by the Japanese Patent office
Commission’s Key Figures 2007. innovation (JPO)7. The gap is still consider-
At EU level, technological speciali- able for 20-year protection. In
sation has not evolved much Larger markets would logically 2004, a European patent exam-
towards R&D intensive industries, result in a higher expected return ined by the EPO and validated in
which explains the lack of ‘visible on investment in R&D. The market 13 member states cost more than
progress’ over the past few years2. size hypothesis may explain why €20,000, against €1,800 in the US
This technological specialisation the US has an above-average R&D and €1,500 in Japan.
2
factor is taken by the Commission intensity (larger than its industrial
“The lack of visible
progress between 2002 to explain both the European R&D structure would suggest). The US
Table 3
and 2005 is largely due ‘inertia’ (the business R&D benefits from a huge and homoge-
to the fact that busi- European patent costs (enforced
ness research expendi-
intensity has been very stable over neous market, with one main lan-
in 13 countries) relative to the US
tures depend on the the past twenty years) and the EU guage and one regulation5. In and Japan
structure of industry, gap with respect to the business Europe, sending a product from
which evolves slowly”, Cumulated
European R&D intensity of the US3. Amsterdam for sale in Brussels is Total cost
fees and
Commissioner still considered an ‘export’, where- for 20
Potocnik, December
translation
However, the results obtained by as in the US a product made in New years (**)
2007, ‘Towards an open costs (*)
and competitive Mathieu and van Pottelsberghe York and sold in Los Angeles is US 11 9
European Research (2008) suggest that when the labelled ‘distribution’. Besides
Area’, in ‘The future of Japan 14 7
Science and Technology technological specialisation of these proverbial examples, a large
in Europe’, MCTES. countries is taken body of evidence Source: Adapted from van Pottelsberghe and
François (2006). These figures represent the
into account, has been published simulated costs of a European patent divided
3
“..The EU/US BERD ‘The ‘average’ EU
deficit cannot be attrib- Sweden and, to a on the lack of by the simulated cost of an average patent in
uted to the fact that lesser extent, the US government has European integra- the US and in Japan. (*) The costs include the
individual European expenses (fees and translation costs) for a
companies perform still display above- actually reduced its tion. And an addi- patent examined by the European Patent
less R&D than their US average R&D support to R&D activity tional key growth Office (EPO) and validated in 13 European
counterparts in the countries after granting. (**) The total cost
intensity. Something ingredient is still
same sector: the main
other than techno-
over the past ten years.’ missing: an EU-wide for 20 years also includes the renewal fees
reason for the deficit is for 20 years in 13 European countries. These
linked to differences logical specialisa- financing solution costs are related to the absolute cost of an
between the European
and American industrial tion thus seems to drive R&D for emerging companies6. average patent. The recently ratified London
Protocol will reduce translation costs
structures.” (Key intensity in Sweden and the US. somewhat.
Figures 2007, p. 35)
The next section puts forward ten- Emblematic of the lack of market
4
In addition to the tative explanations for the US and integration is the way the These costs only include the filing
numerous innovation- Swedish exceptions and draws innovation system works in examination, validation, transla-
related policy recom-
mendations proposed lessons for EU and national policy. Europe. The European patent tion and renewal fees. They do not
by expert groups, such system, and hence the European reflect the managerial complexity
as the Aho Group report
(2006) and the second
3. HOW CAN EUROPE STIMULATE market for technology, is highly of enforcing patent portfolios in
report of the Knowledge BUSINESS R&D? fragmented. Once a patent has several European countries, nor do
for Growth Group been granted by the European they include the litigation costs in
(2007) mentioned in
section 2 of this Policy One important driver of business Patent Office (EPO), it must be case of infringement. The policy
Brief. R&D expenditure is the expected validated, translated, monitored implication is straightforward. The
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

failure to create an EU patent

places a heavy burden on the
shoulders of European innovators
Figure 5: Business-funded R&D and R&D carried out in institutions of
higher education as a percentage of GDP, 2006 or closest 07

bruegelpolicybrief
and entrepreneurs at the very 3

beginning of the innovation SE

process - a clear comparative dis- 2.5
FI

Industry-financed R&D, % of GDP

advantage for Europe with respect
to the US and Japan. 2
DE

1.5
More and better academic DK
LU
research FR BE
AT
1
CZ NL
IE
SL
Market size may explain US per- ES
UK
0.5
formance with regard to R&D HU IT
RO PL PT
intensity, but it does not explain SL GR
0
the performance of Sweden. The 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
explanation here is probably linked R&D performed by the higher education sector, % of GDP
to the relatively very high level of
spending on academic research, Source: OECD, MSTI, 2007.
the highest (as a percentage of Not only does academic research expect to increase their technical 5
The idea that there is a
positive relationship
GDP) in the whole OECD area, as feed ideas to the market, but it staff in China and India, while they between the size of a
illustrated by Figure 5. This strong also attracts more funding from anticipate a substantial decrease country and its propen-
emphasis on academic research is the business sector and promotes in such staff in Europe. sity to invest in R&D is
empirically and theoret-
also a stimulus for business R&D: the setting up of scientific clus- ically supported by
universities generate new ideas ters. For instance, The important role Guellec (1999) and
Desmet and Parente
which are then transferred to the Abramovsky et al played by academic (2006).
private sector. The transformation (2007) show that, ‘High academic research as provider
6
of these ideas into products or in the UK, research spend of ideas to the busi- See Philippon and
Véron (2008), Bruegel
processes requires further applied universities with a ness sector and as Policy Brief 2008/1
correlates with high
research activity and high scientific out- a driver of foreign ‘Financing Europe’s fast
development. Not surprisingly, the put attract signifi- business R&D spend.’ R&D expenditure movers’.

four countries in Figure 5 with the cantly more local implies a need for 7
See van Pottelsberghe
highest academic R&D intensities and foreign research relatively more and François (2006).
are also the four countries with the laboratories to their neighbour- resources to be devoted to higher 8
Guellec and van
highest business R&D intensities. hood. This question is key because education research activities. Pottelsberghe (2004)
provide evidence sug-
Provided effective technology gaining a technological edge is the Indeed the recent Bruegel Policy gesting that the social
transfer systems are main driving force Brief on European universities return to academic
put in place, academic ‘The European behind foreign busi- underlines that Europe invests too research is higher than
the social return to
research is probably patent system, and ness R&D investment, little in higher education and that business R&D.
the most effective be it in the US, in ‘European universities suffer from
source of new ideas, hence the market
9
Walsh (2007) docu-
Europe, or elsewhere. poor governance, insufficient ments evidence on
which in turn induce for technology, is In fact, large firms autonomy and often perverse more than 750 foreign-
further research in the highly fragmented.’ nowadays increasing- incentives’10. In addition to reme- owned R&D centres in
China in 2005.
business sector8. In ly invest in emerging dying these three failings, govern-
this respect, the European markets, which provide a high- ments should also provide more 10
Philippe Aghion,
André Sapir, Mathias
Research Council (ERC), which quality labour force at much lower funding for universities’ research Dewatripont, Caroline
provides merit-based fundamental cost than in Europe9. As shown by activities. The alternative for Hoxby and Andreu Mas-
research grants, is a recent Thursby and Thursby (2006) in Europe will be to lose related busi- Colell, Bruegel Policy
Brief, 2007/4, ‘Why
positive example of what the EU their survey of US and European ness research, and ultimately to Reform Europe's
can achieve. firms, a majority of respondents lose business. Universities?’.
 EUROPE’S R&D: MISSING THE WRONG TARGETS?

REFERENCES:

08 Abramovsky L., R. Harrison and H. Simpson, 2007, University Research and the Location of Business R&D, The Economic
bruegelpolicybrief

Journal, 117 (519), pp. 114-141.

Aho Group, 2006, Creating an Innovative Europe, Report of the Independent Expert Group on R&D and Innovation appoint-
ed following the Hampton Court Summit, available at http://europa.eu.int/invest-in-research/. Last accessed 25 January
2008.

Desmet K. and S. L. Parente (2006), Bigger is better: market size, demand elasticity and resistance to technology adop-
tion, CEPR Discussion Paper, 5825, September, 36p.

Guellec D. (1999), A la recherche du tant perdu, Revue Française d’Économie, 14(1), pp. 117-169.

Guellec D. and B. van Pottelsberghe de la Potterie (2004), From R&D to productivity growth: do the institutional settings
and the sources of funds of R&D matter? Oxford Bulletin of Economics and Statistics, 66(3), pp. 353-376.

European Commission 2007, Key Figures 2007 – Towards a European Research Area – Science, Technology and
Innovation.

Knowledge for Growth Group, Report on the EU’s R&D Deficit & Innovation Policy, 2007 http://ec.europa.eu/invest-in-
research/pdf/download_en/rdd_deficit_report0207.pdf; (Rapporteur: Mary O’Sullivan). Last accessed 25 January 2008.

Mathieu A. and B. van Pottelsberghe de la Potterie, 2008, A note on the drivers of R&D intensity, CEPR Discussion Paper
6684.

Thursby J. and M. Thursby, 2006, Here or there? A survey of the factors in multinational R&D location, National Research
Council of the National Academies, Washington D.C.

van Pottelsberghe de la Potterie B. and D. François, 2006, The cost factor in patent systems, CEPR Discussion Paper 5944.

Walsh, K.A., 2007, China R&D: a high-tech field of dreams, Asia Pacific Business Review, 13(3), 321-365.

The author thanks all those who reviewed this Policy Brief and Jérémie Cohen-Setton at Bruegel for excellent
research assistance.

Bruegel is a European think tank devoted to international economics, which started operations in Brussels in 2005.
It is supported by European governments and international corporations. Bruegel’s aim is to contribute to the quality
of economic policymaking in Europe through open, fact-based and policy-relevant research, analysis and discussion.

© Bruegel 2008. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted in the original
language without explicit permission provided that the source is acknowledged. The Bruegel Policy Brief Series is pub-
lished under the editorial responsibility of Jean Pisani-Ferry, Director. Opinions expressed in this publication are those of
the author(s) alone.

Visit www.bruegel.org for information on Bruegel's activities and publications.

Bruegel - Rue de la Charité 33, B-1210 Brussels - phone (+32) 2 227 4210 info@bruegel.org