вторник, 30 сентября 2014 г.

25 years after the fall: indicators of postcommunist science

в сокращенном варианте текст опубликован в HERB #2 (pdf)

Ivan SterligovAlfia Enikeeva 

In 1985 COMECON countries published more than 64 thousand documents indexed in the US-produced Web of Science database, of which only 1-2% were coauthored with American scientists. USSR alone accounted for almost 40 thousand publications, or 4,4% of the total world output, not counting numerous quality Soviet journals missing from the Web of Science. By 2013 Russia’s share was only 1,6% and all former partner communist states switched to EU or USA as their main collaborators. This article aims to provide a brief statistical overview of these massive changes by combining bibliometric data with some general development indicators and historical remarks.

Communist scientific system at its peak

By the end of the 1980-s R&D system in the USSR was firmly established as the second largest in the world with more than 1,5 million researchers (including university lecturers) and gross expenditures amounting to 1,5% of GDP. The largest part of it was not basic but applied, with thousands of research centers and design bureaus across the Union, usually controlled by sectorial ministries. Basic research was done mostly in 330 institutes of Academy of Sciences and several prominent institutes housing megascience facilities operated by ministry of nuclear science and technology. Medical and agricultural sciences were branched into separate Academies, the first one having 79 research centers and the latter more than 100. Besides that all soviet republics had their own “lesser” academies with total number of researchers just slightly behind that of “major” nationwide Academy. Overall number of researchers in all academies of the USSR by 1988 was 150 thousand. University research was very limited comparing to the US, its share in soviet R&D by 1990 was estimated to be ca. 10%.

Similar organizational models were actively promoted in virtually all countries under soviet influence, which led to establishment of Academies of Sciences in all COMECON countries, Yugoslavia, Albania, China and North Korea. Together these Academies had hundreds of research institutes active in all branches of modern science and humanities. However, those countries that had an established tradition of university research (notably Poland and Czechoslovakia) were fully allowed to foster their HEIs alongside academic research centers.

Degree of cooperation between COMECON  researchers and organizations was varied. Ambitious integration project called “Comprehensive Program for Scientific and Technical Progress up to the Year 2000” was adopted only in December 1985, near the end of the Soviet era. With 93 projects and 800 subprojects within 5 broad priority areas it was a centrally planned analogue of European Framework programs. Each project was led by a soviet institute, which awarded R&D contracts to COMECON partner organizations. International coauthorship between Eastern Bloc and between Communist and Capitalist states clearly did exist, but we can’t correctly estimate it with WoS data because of frequent omissions of affiliations information in this database prior to 1990-s.

Overall coverage of soviet science by western citation databases remained very limited until 1983-1985, when Philadelphia-based Institute for Scientific Information added many new soviet journals to their Web of Science database (WoS), leading to a massive increase of publication counts for USSR and its allies. In 1989 USSR had published 40,823  WoS documents, only surpassed by the US, UK and Japan. Poland had 6,326, Czechoslovakia 4,859 - both ahead of Austria, Finland or Norway. Top disciplines for COMECON countries were physics and chemistry, which contrasted with both US and Japan (biochemistry) and UK (medicine).

After the fall

Demise of the USSR, collapse of Eastern European communist regimes and removal of the Iron Curtain in 1988-1991 led to several drastic consequences for R&D. Funding was severely reduced, priorities shifted, and new national systems emerged free from Soviet influence. All countries reacted very differently.

Figure 1. Total number of publications in the Web of Science, 1993-2013 (all document types and all indices)

Russian gross expenditures on R&D (GERD) dropped more than fourfold in 1990-1992, GERD/GDP ratio plummeted from 1,43% in 1991 to 0,85% in 1995. Lack of demand for new Russian technologies led to prolonged crisis of applied research centers.  Scientific personnel shrank from 130 researchers per 10.000 workforce in 1990 to 60 in 1995. By 2012 we still have almost two times less GERD than in 1990 (in constant prices) and Russia’s GERD/GDP ratio is 1,12%. Nevertheless, Russia saw a marked increase in government spending on R&D in 2000s and rapid growth in number of PhD students, PhD holders and universities. Sadly, these statistical achievements had little effect on publication counts (see fig.1) and on total number of employed researchers, which dropped from 425 thousand in 2000 to 372 thousand in 2012 (not including university lecturers).

Figure 2. GERD in 1993-2013 for several former COMECON countries (million 2005 dollars - constant prices and PPP. Data source: OECD)

A number of former Eastern Bloc countries suffered similar decline but had much more success in revitalizing their R&D. Poland and Czech Republic turned out to be the most active among bigger states both in terms of research expenditures and publication counts (see fig.  1-2), with Kazakhstan, Vietnam and Baltic states also showing promising growth (see tab. 1). For example, number of researchers in Kazakhstan has increased 50% in 2009-2013, and Estonia’s ratio of GERD to GDP reached record 2,37% in 2011. Vietnam has seen a rapid increase in WoS publications in line with its industry-driven economic growth and strong ties with other rapidly developing Asian countries.

Other CIS countries achieve more modest results. Don’t let recent growth in publication counts fool you – they are in large part provided by physicists working in CERN or EU\US-based collaborations which now produce thousands of articles in high-ranking journals per year.[1] On the whole, however, basic research in Armenia, Georgia, Tajikistan, Turkmanistan, Uzbekistan and several other states is almost completely strangled by lack of systemic funding, and there’s no prominent improvement there in recent years.

Shifts in international collaboration, institutional origin and themes of publications

Table 1. Main indications of research in ex-COMECON countries 

Contrary to popular misconception, division between universities doing mostly teaching and institutes doing mostly research isn’t something entirely soviet and outdated. Germany serves as the best example of an R&D system spearheaded by non-graduate institutions, with only 1 Nobel laureate out of 9 in physics, medicine and chemistry in 1990-2013 coming from a university, and zero universities in top-40 in QS or THE. Germany also is an example of the most radical, quick and effective reform of an ex-communist Academy of sciences. In 1991-1993 all the institutes of GDR Academy were evaluated, then many of them were closed and the rest formed Leibniz-Gemeinschaft modeled after Max-Planck-Gesellschaft and Fraunhofer-Gesellschaft. Now it employs ca. 17 thousand people working in 87 organizations.

The fate of other academies is different. Russian academies  amidst decrease of funding and massive brain drain managed to markedly increase number of research institutes and academicians before being forced into reform, the outcomes of which are yet to be seen. Russian Academy of Sciences’ resilience and opposition to changes is remarkable among other postcommunist countries. However, almost all other academies - Bulgarian, Polish, Ukrainian, Armenian etc. - do exist and run their networks of institutes, but the numbers of scientists there are on decline. Partly it’s because governments are pursuing American “Triple Helix” agenda and tend to provide much more money to leading universities. Our data shows a clear shift towards universities as main force behind WoS-indexed papers almost everywhere, but shares of universities among ex-COMECON countries vary very significantly.

Another trend is an almost universal turn to USA and EU15 countries for collaboration and coauthorship. Russia is maintaining its role as a priority partner only for ex-USSR states, but even Belarus and Armenia, our closest allies, now have more papers coauthored with EU15 than with Russia.  Needless to say, new EU-member states are receiving huge benefits by access to Framework Programs, Horizon 2020 and European Research Council grants. Not only they provide much-needed money (€427 million in FP7 for Poland alone), they do it in a clear and competitive way and actively foster wider collaboration.

To sum up, there are two distinct patterns among former COMECON members:

1.      -  EU member states with growing publication counts, high rates of collaboration with EU15 countries, high shares of universities (~70-75%) and increasing priority for medical research popular in developed capitalist countries, at the same time focusing on relatively new areas (ICT in Estonia as a most successful example). Their integration in European research area is definitely on the way.

2.      -  Russia, Ukraine and Belarus continue to pursue a more conservative path, with only half of publications authored by universities’ employees, and preserved dominance of physics as “the” science. Rate of collaboration with EU and US peers in these countries is also high, but there’s no comparable growth of publication output.

Eastern and southern ex-USSR states receiving no direct EU support have to rely on their own.[2] For most of them, continued financial struggles have meant widespread brain drain of USSR-trained researchers, which makes current R&D capacity building very complicated. The country with most marked improvements is Kazakhstan, with its strong commitment to creating national system of research universities that really works. On the whole, however, share of non-EU ex-communist states in world scientific production has yet to reach USSR levels.

[1] It is worth mentioning that in 2011-2013 LHC and other large-scale collaborations in high energy and particle physics have severely skewed bibliometric indicators for many countries because of extraordinary high number of resulting papers, and authors per paper (at times more than 3000 authors per article). Thus one physicist form a small country included in, say, ATLAS and CMS collaborations can provide that country and his institute with 100 or even 200 articles in top physics journals each year. The percentage of physics collaboration papers in all papers published in 2013 is 11,4% for Belarus, 15,4% for Azerbaijan, 22,4% for Armenia and 24,7% for Georgia (for other countries in tab. 1 this share is less than 5%, according to our lower eastimate).
[2] Perhaps the only exception is Azerbaijan with its revived ties with Turkey becoming stronger each year.

Комментариев нет:

Отправить комментарий