Académique Documents
Professionnel Documents
Culture Documents
SEPTEMBER 2007
bruegelpolicybrief
WHY REFORM
EUROPE’S UNIVERSITIES?
by Philippe Aghion SUMMARY Recently published international rankings indicate that the perfor-
Non-Resident Senior Fellow at Bruegel
and Professor of Economics at
mance gap between European and American universities is large and, in
Harvard University particular, that the best European universities lag far behind the best American
philippe.aghion@bruegel.org
universities. The country performance index we construct using the Shanghai
Mathias Dewatripont ranking confirms that, despite the good performance of some countries,
Professor of Economics at
Université Libre de Bruxelles
Europe as a whole trails the US by a wide margin. The reason for this situation,
which contributes to Europe’s lagging growth performance, is two-fold. First,
Caroline Hoxby Europe invests too little in higher education. Total public and private spending
Professor of Economics at
Harvard University on higher education in EU25 accounts for barely 1.3% of GDP, against 3.3% in
Andreu Mas-Colell the US. This translates into average spending of less than €10,000 per student
Professor of Economics at in EU25 versus more than €35,000 in the US. Second, European universities
Universitat Pompeu Fabra
suffer from poor governance, insufficient autonomy and often perverse
André Sapir incentives. Our own survey of European universities shows that both factors
Senior Fellow at Bruegel
and Professor of Economics at
contribute to the EU’s poor performance and that reform should take place on
Université Libre de Bruxelles both fronts, because autonomy also increases the efficiency of spending.
andre.sapir@bruegel.org
POLICY CHALLENGE
The EU-US performance gap for
Shanghai Top 100 universities
(US=100)
If Europe is to be a leader in the global
100 knowledge economy – and if universities
are to produce the top-level research
needed to achieve this – comprehensive
reform of higher education is the order
of the day. While there is more than one
university system that works, most
50
European countries should invest an
extra one percent of their GDP in higher
education each year. The extra funding
may be public or private. To make this
additional funding ‘bite’, universities
0 should be given more autonomy in
EU15 EU25 US
budgets, hiring and remuneration.
Source: Bruegel country
performance index, see Table 1.
WHY REFORM EUROPE’S UNIVERSITIES?
EUROPEAN growth has been disap- makers aware of the magnitude of by the publication, since 2003, of
02 pointing for the last 30 years but
policymakers have only recently
the problem and sparked a public
debate on university reform.
the Shanghai ranking which mea-
sures university research perfor-
bruegelpolicybrief
bruegelpolicybrief
100, 200 and 500, and the last one Country
(millions) Top 50 Top 100 Top 200 Top 500
receives a score of 1. There are, Austria 8 0 0 0 53
obviously, fewer zero entries in a Belgium 10 0 0 61 122
column as one moves from the Top
Czech Republic 10 0 0 0 13
50 to the Top 500 as it is easier for
Denmark 5 0 75 114 161
a country to have universities
Finland 5 0 46 75 81
appearing in the Top 500 than in
France 60 3 15 29 45
the Top 50.
Germany 83 0 17 37 67
Greece 11 0 0 0 12
Table 1 reveals several interesting
Hungary 10 0 0 0 13
findings:
Ireland 4 0 0 0 50
Italy 58 0 0 11 34
• First, the United States comple-
Netherlands 16 20 51 76 131
tely dominates all European coun-
Poland 38 0 0 0 4
tries in the Top 50 universities.
Only Switzerland and the United Spain 43 0 0 0 14
Kingdom rival the US on a per Sweden 9 7 117 179 217
capita basis. By contrast, the EU15 UK 60 72 86 98 124
and EU25, with a greater popula- EU15 383 13 26 41 67
tion than the US, score much lower. EU25 487 10 21 32 54
• Second, the top 4 US states Australia 20 0 31 66 101
(Massachusetts, California, New Canada 32 39 54 63 104
York and Pennsylvania) score bet- Japan 128 14 17 24 27
ter than any European state in the Norway 5 0 66 91 107
Top 50 and Top 100. Switzerland 7 97 166 228 230
• Third, country performance gaps US 294 100 100 100 100
close as one enlarges the number California 36 234 199 163 103
of universities considered. In parti- Massachusetts 6 449 308 302 263
cular, the gap between the EU15 or New York 19 196 167 139 148
the EU25 and the US narrows as
Pennsylvania 12 111 177 161 115
one moves from the Top 50 to the
Texas 23 33 61 83 103
Top 500. In part this is due to the
way the scores are constructed, Source: Bruegel computation based on the Shanghai Jiao Tong University Academic Ranking of
World Universities (2006).
but it mostly reflects a reality:
Note: All figures are rounded.
American universities dominate
European universities in the top Top 200 and Top 500. In contrast, • Richer countries spend relatively
tier (the Top 50 and Top 100), but southern and eastern Europe lag far more on higher education than poo-
Europe has many good universi- behind. France and Germany do rer countries.
ties in the second (the next 100) relatively poorly, except in the third • The US spends a lot more on higher
and the third (the next 300) tiers. tier, the universities ranked bet- education than any European coun-
• Fourth, there are major differences ween 301 and 500. try, especially thanks to private fun-
among European countries: ding. But public spending alone is
Switzerland, the UK and Sweden do 2. WHAT EXPLAINS RESEARCH relatively higher than in the EU.
particularly well, even in the Top PERFORMANCE IN EUROPE? • The Nordic countries also spend a
100, where they outperform lot, with most of the money coming
(Switzerland and Sweden) or An obvious starting point for eco- from public sources.
almost match (the UK) the United nomists is money. Table 2 pre- • The UK spends surprisingly little
States on a per capita basis. Other sents aggregate data on the levels (more on this later).
Nordic countries (Denmark and of private and public expenditure
Finland), Belgium and the on higher education across coun- There is a strong positive correlation
Netherlands also do fairly well in the tries. The main findings are that: between expenditure per student
WHY REFORM EUROPE’S UNIVERSITIES?
Country Public Private Total Public Private Total dents at least two3. This left us with
a total sample of 66 universities,
Austria 1.4 0.1 1.5 11.0 0.5 11.5
with an average response rate of
Belgium 1.4 0.2 1.6 10.6 1.6 12.2
41 percent for the ten countries
Czech Republic 0.8 0.1 0.9 2.3 0.4 2.7
considered. We were able to check
Denmark 2.7 0.0 2.7 25.6 0.4 26.0
that, for each country, respondent
Finland 2.1 0.1 2.2 10.3 0.3 10.6
universities have an average
France 1.0 0.2 1.2 7.5 1.2 8.7 Shanghai 500 rank fairly close to
Germany 1.1 0.1 1.2 11.5 0.9 12.4 that of the whole population of uni-
Greece 1.2 0.0 1.2 3.3 0.0 3.3 versities from that country, so that
Hungary 1.1 0.3 1.4 2.6 0.6 3.2 we could be satisfied of the repre-
Ireland 1.2 0.2 1.4 9.7 1.6 11.3 sentativity of our sample4.
Italy 0.8 0.2 1.0 5.6 1.4 7.0
Netherlands 1.3 0.3 1.6 13.0 2.7 15.7 Table 3 provides country averages
Poland 1.1 -* -* 1.7 -* -* on a variety of dimensions5. It
Spain 1.0 0.3 1.3 4.0 1.2 5.2 confirms the high degree of hete-
Sweden 2.1 0.2 2.3 18.9 1.8 20.7 rogeneity between countries for
UK 0.8 0.3 1.1 8.4 3.1 11.5 the universities in the Top 500:
EU25 1.1 0.2 1.3 7.3 1.4 8.7
•Southern European countries
Japan 0.5 0.6 1.1 6.5 7.3 13.8 (Italy and Spain) have very large
2
The 2006 Shanghai (more than 40 thousand students
US 1.5 1.8 3.3 16.6 19.9 36.5
ranking includes
roughly 200 on average) but not well-funded
European universi- Source: European Commission, DG Research, for expenditure as percent of GDP. Figures in universities.
ties belonging to euros per student were obtained by applying these percentages to GDP and dividing by the • Sweden and the Netherlands
the EU25 and number of full-time equivalent students.
Switzerland. * Not available. Note: Not PPP converted.
have universities of average size
(20-25 thousand students), and
3
The ten countries are: and country performance (measu- overall response rate of 36 percent, better funded.
Belgium (4 responses red by the Top 500 performance which can be considered very •The UK and Switzerland have
out of 7 universities in
the Shanghai 500 ran- values in Table 1), as shown in satisfactory. We decided to focus small (10-15 thousand students)
king), Denmark (2 out Figure 1.
of 5), Germany (11 out
of 40), Ireland (2 out of Figure 1: Relationship between expenditure
3), Italy (9 out of 23), However, these aggregate data do
Netherlands (4 out of not indicate how the money is split per student and country performance
12), Spain (6 out of 9), between higher education institu- Country performance index
Sweden (5 out of 11), 250
Switzerland (6 out of 8)
tions, in particular between
and the UK (17 out of research-oriented and teaching- Sweden
43). We left out France, oriented universities. In the 200
because only 4 out of remainder of this section we there-
21 universities respon-
ded and, moreover, uni- fore present the results of a survey Denmark
150
versity budgetary data questionnaire which elicits infor-
Netherlands
are not comparable mation on individual budgets and UK Belgium
with those of other
countries. on the governance of top research 100 US
performers. Finland
4
In fact, respondents Germany
Austria
had a somewhat higher 50
France Ireland
rank for all countries
A survey of European universities Italy
Age
Table 3: Characteristics of the universities in the sample (country averages)
Number Budget per
Public Budget Building Hiring Wage-setting
Faculty with
05
bruegelpolicybrief
Country of students student in-house
(years) (thousands) (1 000€)* status+ § § §
autonomy ownership autonomy autonomy §
PhD (%)
Belgium 284 21.7 11.3 0.5 0.4 1.0 1.0 0.0 63
Denmark 59 18.2 11.4 1.0 1.0 0.3 0.5 0.5 40
Germany 289 26.2 9.6 0.9 0.0 0.5 0.8 0.0 40
Ireland 259 16.3 12.7 0.5 0.5 1.0 1.0 0.0 49
Italy 444 44.9 10.1 1.0 0.9 1.0 0.4 0.0 24
Netherlands 217 21.4 20.5 0.8 0.8 1.0 0.8 0.2 33
Spain 342 44.8 7.0 1.0 0.5 1.0 0.5 0.0 69
Sweden 266 27.1 16.2 0.8 0.8 0.2 1.0 1.0 58
Switzerland 326 12.8 26.2 0.8 0.1 0.4 0.8 0.0 24
UK 242 14.6 24.5 0.5 0.9 0.9 1.0 0.8 8
Total 290 24.9 16.1 0.8 0.6 0.8 0.8 0.3 29
Source: Bruegel survey.
* PPP adjusted. + 1 if public, 0 if private. § 1 if yes, 0 if no.
and very well funded universities. open to hiring scholars with PhDs accumulated knowledge, but may
Comparing with the aggregate from other institutions), and small also be complicated to reform.
information on expenditure in in large countries (Germany, Italy
Figure 1, one observes that the UK and the UK, but not in Spain). This Preliminary evidence
significantly favours top research finding clearly reflects the absence
performers since the universities of significant academic mobility Our survey allows us to examine
in our sample (which belong to the between European countries. how budget per student and
group of top universities) have a various measures of university
budget per student about twice as A striking fact is thus the high governance correlate with
large as the average for all univer- variance in university governance research performance as measu-
sities in the country. across European countries, even red by the Shanghai ranking. Table
among those which are performing 4 shows that the research perfor-
There is also a great deal of hetero- well in terms of research. For exam- mance of a university is:
geneity – albeit with some general ple, among the three European
trends – as far as university gover- countries with the best perfor- • positively correlated with the size
nance is concerned: mance index, endogamy is high in of its budget per student: the higher
Sweden but low in Switzerland and
• State intervention is clearly per- the UK, and universities are mostly Table 4: Correlation between
vasive, even when universities are public in Denmark, Sweden and budget and university governance,
not public. Switzerland whereas they are and research performance*
• Wage-setting autonomy is rare, mostly private in the Netherlands
with Sweden and the UK being the and the UK. Correlation
Characteristics
prime exceptions. coefficient
• Building ownership by the univer- One dimension where there is little Budget per student +0.61
sity is commonplace (except in the variance across European coun- University governance:
Nordic countries and Switzerland). tries is the age of universities. Top Public status+ -0.35
• Hiring autonomy is prevalent, European universities are old insti- Budget autonomy§ +0.16
except in southern Europe. tutions: the average age of the 66 Building autonomy§ -0.01
• Endogamy (measured as the per- universities in our sample is nearly Hiring autonomy§ +0.20
centage of faculty trained in house 300 years. It ranges from 220 Wage setting autonomy§ +0.27
at PhD level) seems to be negatively years in the Netherlands to 450 Faculty with -0.08
correlated with country size: it is years in Italy. The only outlier is in-house PhD
high in small countries (Belgium, Denmark where the average age is * Measured by the (logarithm of the)
Denmark, Ireland and Sweden, but only 60 years. This suggests that Shanghai ranking
not in Switzerland which is highly European universities have much +
1 if public, 0 if private. § 1 if yes, 0 if no.
WHY REFORM EUROPE’S UNIVERSITIES?
the budget per student, the better We are interested in the effect of
06 the performance;
• negatively correlated with its
budget and university governance
on university research perfor-
Table 5: Effect of budget
and autonomy on
research performance*
bruegelpolicybrief
while their neighbours in Ohio out of our investigation. well with mostly public universities,
enjoy high autonomy. These diffe-
rences are persistent over time
First, the performance gap bet-
ween Europe and America is large,
while the UK also performs well
with a higher share of private uni- 07
bruegelpolicybrief
and often go back to the idiosyn- in particular for the best-perfor- versities, but also higher tuition
cratic origin of American universi- ming universities. fees and a higher degree of stu-
ties, which in turn reflect diffe- dent selection. The UK, however,
rences in the preferences of univer- Second, as we broaden the investi- differs significantly from
sity founders (eg Benjamin gation from the Top 50 to the Top Switzerland and Sweden in one
Franklin founded the private 500 universities in the Shanghai respect. All three perform very well
University of Pennsylvania, whe- ranking, the relative performance in the top tier (Top 50 and Top
reas Thomas Jefferson was the of European countries improves 100), but the UK performs relati-
founder of the public University of compared to the US. This, in turn, vely less well in the remainder of
Virginia). suggests strongly that quality the Top 500. This is due to the fact
variance is lower among European that the UK concentrates its less
Our strategy is to take US states' universities than among their than average higher education
differences in university autonomy American counterparts. It also sug- budget (in terms of GDP) heavily
as given and then ask the following gests that what Europe lacks most on its top institutions.
question: Does a given investment is top-class universities.
in higher education produce more Indeed, a fourth lesson is that
patenting in a US state if universi- Third, there is more than one model money helps performance.
ties in that state are more autono- of university system that appears
mous? The details of the statistical to work. For example, both Fifth, autonomy is good for
test are reported in Box 2. The ans- Switzerland and Sweden are doing research performance.
wer to our question is a resounding
‘yes’. As illustrated in Figure 2, the
effect of additional spending on Box 2: University funding, autonomy and innovation:
patenting is roughly twice as high data and methodology
Data
for states with more university
autonomy. Autonomy therefore For research expenditure, we use the detailed data in Aghion et al (2007)6 on how
greatly enhances the efficiency of much each state spent on each type of education in all years from 1947 to 2004. We
spending. This result confirms and know in particular from these time series how much each US state spent on a given
cohort of individuals (eg born in year X) in each year. Thus we know how much was
nicely complements the one from
spent on average on each individual at every stage of his or her studies (from primary
Section 2. school to post-graduate college).
4. CONCLUSIONS For governance, we consider two alternative measures of university autonomy at the
state level: (i) the percentage of universities that are private, keeping in mind that pri-
vate universities are, on average, more autonomous than public universities; (ii) an
In this brief we have investigated aggregate autonomy index for public universities, which is constructed on the basis of
the relationship between univer- several component factors. This index takes the maximum value when the public uni-
sity governance and funding on versities in the state: (a) set their own faculty salaries; (b) set their own tuition fees;
the one hand and various mea- (c) have lump-sum budgeting (as opposed to line item budgeting); (d) can shift funds
sures of performance on the other among major expenditure categories; (e) retain and control tuition revenue and/or
hand. In the first section we have grants; (f) have no ceiling on external faculty positions (and therefore need not hire
faculty internally); (g) have no ceiling on external non-faculty positions (administra-
tried to link our Country tors or technicians); (h) have freedom from pre-audits of their expenditure; (i) can
Performance Index based on the carry over year-end balances (rather than returning them to the state). It emerges
Shanghai ranking of universities to that, as in the case of European universities, the most statistically important compo-
different aspects of university nent factor of this aggregate index is budget autonomy.
6
governance drawn from a survey Philippe Aghion,
Statistical test Leah Boustan, Caroline
questionnaire. In the second sec- Hoxby and Jerome
tion of the brief we have assessed We examine the effect on patenting in a US state of increasing research education fun- Vandenbussche,
how university autonomy affects ding by $1,000 per year and per person over a sustained period, respectively in states Exploiting States’
with highly autonomous universities and in states with less autonomous universities. Mistakes to Evaluate
the patenting impact of university
the Impact of Higher
research funding. Figure 2 illustrates a key result from our test: states with highly autonomous universi-
Education on Growth,
ties enjoy an accumulated impact of research education funding on innovation which is mimeo, Harvard
Several interesting findings come roughly twice as high as that enjoyed by states with less autonomous universities. (2007).
WHY REFORM EUROPE’S UNIVERSITIES?
increases the extent to which addi- Effects on patents per million people
tional research funding improves 0.4
performance.
States with high autonomy universities
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 2007. 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 published under the editorial responsibility
of Jean Pisani-Ferry, Director. Opinions expressed in this publication are those of the author(s) alone.