IV. INTERNATIONAL COMPARISONS
The Panel examined numerous categories of data in trying to benchmark
the activity and health of U.S. mathematical sciences in relation to those
of other countries. It found a number of indicators (see Appendices) and
qualitative observations to be significant, as discussed below.
In comparing the United States with Europe, the Panel found the United
States mathematics produce more research papers. Pacific Rim nations trail
that of the United States and Europe, but have increased their output
significantly since the early 1980s (not shown)
(see Endnote 11).
Publications in the Mathematical Sciences
1989 1991 1992
United States 39.9% 42.1% 38.9% United Kingdom 5.7% 6.3% 5.9% Germany 6.6% 7.0% 6.5% France 7.5% 4.6% 8.5% Other Western Europe 12.7% 12.6% 13.1% Japan 4.3% 4.6% 3.6% Other Pacific Rim 3.9% 3.7% 4.4%
In a separate study by CHI Research, Inc. of ISI journals with emphasis
on research publications in pure mathematics, the proportion of papers published
by various nations remained stable, with U.S. mathematicians authoring 40-50%.
The study noted that international coauthorship was increasing.
Data on the number of Ph.D.s in the mathematical sciences is difficult
to determine by geographic region. More generally, for the natural sciences,
in 1992 there were 6593 doctorates in Asia, 18,951 doctorates in Europe,
and 13,344 doctorates in North America (with the United States producing
12,555 doctorates). However, the report noted, "a declining pool of
college-age students in Europe has not resulted in declining numbers of
natural science and engineering degrees, as has occurred in the United States."
(see Endnote 12)
International Congress Participation
Data on invited speakers at international conferences vary somewhat.
In 1994, 50% of invited one-hour speakers at the quadrennial International
Congress of Mathematicians were from the United States. In the forthcoming
1998 Congress, 38% of the invited one-hour speakers are from the United
States and 48% are from Europe; on the other hand, 48% of the 45-minute
speakers are from the United States (1/3 of them having non-U.S. origins)
and 36% are from Europe. U.S. speakers accounted for 35% of the plenary
speakers at the 1995 International Congress of Industrial and Applied Mathematics.
A tally of the leading awards in mathematics provides another useful
benchmark. Of the 16 mathematicians awarded the Fields Medal between 1970
and 1990, eight resided in the United States at the time of the award and
11 currently do. Four of the eight medalists honored in the 1990s reside
in the United States, but only one was born in the United States. (The Nobel
Prize is not awarded in mathematics; one U.S. mathematician has received
the Nobel Prize in economics.) Of the mathematicians who have received the
Wolf Prize, an award for distinguished scientists, 14 of 27 are U.S. mathematicians
and two others have spent substantial time in the United States.
The Panel, on the basis of its own expertise, undertook a qualitative
benchmarking of subdisciplines, which appears in full as Appendix 2. The
Panel concluded that the United States has strengths in all subdisciplines,
but that it is not the major contributor in some. In several fundamental
subdisciplines, including Foundations, Symbolic Computation, and Ordinary
Differential Equations, foreign contributions outweigh those of the United
States. Overall, however, the U.S. mathematical research enterprise is judged
capable of responding to advances occurring anywhere in the world, an ability
enhanced by the very high level of interaction among mathematicians worldwide.
The Panel was unable to provide meaningful country-by-country comparisons
of research funding. One reason for this is the wide variations in budgetary
and institutional relations between governments and universities. For example,
in Canada, Europe, and Japan, faculty are paid on a 12-month basis for both
teaching and research. This means that all faculty in these countries have
funding for summer research; in the United States, only 35% of active researchers
do. This "have" and "have-not" situation is very destructive
to the fabric of the U.S. mathematical sciences community and decidedly
discouraging to young U.S. researchers. Funding agencies in Europe and Japan
play differing roles from that of NSF or the DoD agencies, often using different
means to promote the health of science and mathematics within their countries.
For example, the Japanese Society for the Promotion of Science spends most
of its money to support university science libraries and visits by distinguished
A few generalizations are possible, however. In Europe, bursaries for
students admitted to graduate school meet their full living costs and do
not need to be supplemented by money earned through teaching. The docent
programs in Germany and Eastern Europe support students completing the doctorate
and intending to seek a post in a research university, as do the collegiate
fellowships in England. Recently, France has introduced postdoctoral fellowships,
as has the European Union. By contrast, the Panel finds that in the United
States, there are few research assistantships for the mathematical sciences.
The United States is overly dependent on teaching funds to provide support
for graduate students in the mathematical sciences, a custom that prolongs
the time to degree and makes the field of mathematics less attractive to
U.S. students. Also, except for a few NSF funded postdoctoral research fellowships
and a few research instructorships funded by universities, postdoctoral
research opportunities for mathematical scientists in the United States
are exceedingly scarce.
In Asia, overall investment in mathematics is rising. The Asian Tiger
countries -- Singapore, Taiwan, Korea, and Hong Kong -- are building strong
research universities and research institutes with significant mathematical
components. Japan has begun a five-year program to double its funding for
basic science. The Chinese National Science Foundation has given the mathematical
sciences its highest priority for development.
Institutes and conference centers are important elements in the infrastructure
supporting the mathematical sciences as important to the field as
are lab facilities and observatories to physics and astronomers. The Institute
for Advanced Study (IAS) in Princeton, N.J. was the first institute to assemble,
for short periods (4-12 months), groups of mathematical science researchers.
Such institutes have become popular in the mathematical sciences and are
viewed as making significant contributions to the advance of the discipline
because they enable explorations of new developments, facilitate collaboration
among mathematical scientists and assist in the sharing of ideas between
mathematical scientists and those of other disciplines. The NSF provides
partial funding to IAS primarily for support of young researchers, core
funding for three other research institutes (MSRI, IMA, NISS) as well as
for DIMACS, a Science and Technology Center that operates much as the other
institutes but concentrates on discrete mathematics, algorithms and theoretical
computer science. Western Europe has six research institutes plus two conference
centers, Canada has three, and the Pacific Rim has several and is planning
more. Germany, which is less than a third the size of the United States,
has two Max Planck Research Institutes, the Oberwohlfach Conference Center
and seven Sonderforschungberich, which are attached to universities and
have some aspects of an institute and some of the U.S. Science and Technology
Centers. Some institutes have permanent faculty and some, e.g. those whose
core funding comes from the NSF, do not. The Centre Nationale des Recherches
Scientifiques (CNRS), in France, and the National Academies of Sciences
in Russia and Eastern Europe have significant numbers of full-time researchers
in the mathematical sciences.
Collaboration with other disciplines
Communication between academic mathematical scientists and other scientists
is poor the world over. Many mathematical scientists have a limited vision
of their capacity to interact with other scientists. Graduate education
is frequently highly specialized. The structure of universities, where decisions
on promotions, awards, and salaries are made by disciplinary departments,
mitigates against collaboration outside ones discipline. The difficult
and time-consuming task of understanding a second discipline is also an
inhibitor. A few U.S. programs, such as the University of Illinois
Beckman Institute, have been created to advance multidisciplinary research,
but few of these efforts have involved mathematical scientists.
Many nations have begun to promote the collaboration of mathematical
scientists in multidisciplinary research. England appears to be doing so
aggressively, especially under the aegis of the Isaac Newton Institute.
In France, there is significant interaction between mathematical scientists
and engineers, in part due to the commonality of their secondary and collegiate
education. Collaborative initiatives are becoming prominent in countries
emphasizing research that drives the economy (e.g., nations of the European
Union and those of the Pacific Rim). Many U.S. mathematicians are becoming
more involved in multidisciplinary research, probably more than those in
other countries, but science and engineering require much more involvement
by U.S. mathematicians. Nonetheless, progress is being made. More than 10%
of the NSF/DMS budget is invested in projects cofunded with other Divisions
and Directorates. The DMS Group Infrastructure Program did provide funding
for collaborative programs, and proposals from U.S. mathematicians have
been well received by the MPS Office of Multidisciplinary Activities and
by the IGERT program. Other recent NSF initiatives promote multidisciplinary
research, often with a strong mathematical component, such as the Knowledge
and Distributed Intelligence (KDI) Initiative begun this fiscal year.
Interactions between academia and industry
Another trend, most apparent in England and the Netherlands, is to foster
interactions between the very different cultures of academic research and
the private sector. England has been especially active on this front with
its Smith Institute and the OCIAM at Oxford. Activity in the Isaac Newton
Institute has led to significant private investment in academic mathematics.
Other European countries have begun such programs, but progress has been
slow; mixing the cultures of academia and industry is not easy, and there
is concern over intellectual property rights and industrial privacy. In
Japan, there is little interaction. In the United States, some mathematics
departments and institutes (notably the Institute for Mathematics and Its
Applications) interact with industrial and financial entities, but they
are very much in the minority. Involvement of the U.S. mathematical sciences
in the NSF GOALI (Grant Opportunities for Academic Liaison with Industry)
program is small compared to that of other sciences, but is beginning to
grow. The Division of Mathematical Sciences (DMS) has recently collaborated
with DARPA (the Defense Advanced Research Projects Agency) to fund several
initiatives that require collaboration by mathematical scientists with other
academic and industrial scientists.
The expectation of an academic career, rather than one in industry, is
particularly strong in the United States. Some 75% of new U.S. doctoral
mathematical scientists anticipate academic positions. In continental Europe,
by contrast, many universities have Diploma programs from which students
seek nonacademic positions. The English universities have recently introduced
a new degree intermediate between the baccalaureate degree and the Ph.D.
which serves that purpose. Many U.S. mathematical science departments have
discussed a "professional masters" degree, which would emulate
the Diploma programs, but few departments have tested or established them.
Although U.S. doctoral programs in the mathematical sciences are extremely
strong, U.S. undergraduate programs offer less exposure to mathematics,
at less depth, than do those in Europe and Asia. There are two important
reasons for this: (i) U.S. undergraduates arrive at the college level with
less knowledge of mathematics: 50% or more are unprepared to begin the calculus;
(ii) In other countries, undergraduate mathematical science students concentrate
entirely on the mathematical sciences and related subjects, while U.S. students
spend at least 50% of their time on unrelated subjects. This has meant that
graduates of U.S. undergraduate programs must spend time catching up with
their European counterparts, extending time to the doctorate. On the other
hand, the U.S. system allows students greater opportunities to explore outside
their discipline and to change specialties. For example, of the students
with A and A+ in high school who enrolled in U.S. undergraduate schools
in the mathematics sciences in 1984, 75% switched to other programs while
61% of those earning a mathematical science degree by 1989 were recruits
from other fields.(see Endnote 13)
Ph.D. recipients from the best universities, whatever the country, are
at the same level of achievement and preparedness for research. U.S. graduate
departments generally offer a wider range of fields in which to specialize
than is the case in Europe or Asia, but recent developments in the European
Union mean that students in Europe can easily move to other universities,
possibly in another country. This allows them greater variety of specialization.
European graduate students also receive funding sufficient to cover their
living expenses. In the United States, most mathematical science students
need to teach to cover their living expenses.
Anecdotal information suggests that a much larger percentage of students
who begin a doctoral degree program in the mathematical sciences in the
United States fail to earn that degree than is the case in Western Europe.
This is especially true for U.S students. A number of U.S. universities
report that by the third year, no U.S. citizen remains in their doctoral
Attractiveness of the field to the young
During the last decade, the number of U.S. citizens pursuing degrees
in the mathematical sciences has suffered a decided decline. Between 1985
and 1995, U.S. freshmen interested in the mathematical sciences declined
by 32%, and by 23% among the top students
(see Endnote 14). This situation is mirrored in
other nations. The Netherlands, Germany, France, Russia, and Poland all
have reported significant losses in mathematics enrollment during the past
five years. In the last three years, there has been a steady decline in
the numbers of applications to U.S. graduate schools in mathematics by Chinese
students, which is probably a sign of diminishing interest in that country.
European trends toward applications and centralization
In the European Union, the European Commission has shifted emphasis in
the direction of mathematical applications which enhance wealth creation
or the quality of life. In the current Fourth Framework Programme, almost
the only support of pure science is budgeted in the Human Capital and Mobility
program. This program does carry benefits for mathematics as a whole in
supporting the movement of postgraduates and postdocs between countries
of the EU, in funding conferences and networks, and in supporting an agenda
that covers a larger proportion of the needs of mathematicians than those
of laboratory-based scientists. However, the shifting of funds from local
to central control and from fundamental research to applications does not
have the universal support of the mathematics community.
On the positive side, the U.S. mathematics community leads other nations
in a large number of subdisciplines and is judged overall to be capable
of responding to breakthroughs occurring elsewhere in any area of mathematics.
Individually, U.S. mathematicians have won more than their share of prestigious
awards in the field.
At the same time, the mathematics community in the United States shares
with other nations significant disciplinary challenges including a condition
of isolation from other fields of science and engineering, a decline in
the number of young people entering the field, and a low level of interaction
with nonacademic fields, especially in the private sector.
The Panel gained the sense that mathematicians in the United States feel
themselves disadvantaged in comparison with mathematicians of other countries,
most notably in public support. This low morale is not evident in Western
Europe or the Pacific Rim. The European Union is expanding opportunities
and funding for graduate students and postdoctorates. U.S. students are
overly dependent on teaching income, which extends time to degree, decreases
the attractiveness of mathematics to younger students, and contributes substantially
to the fragility of the U.S. mathematical enterprise.