SPATIO-TEMPORAL PATTERNS OF KNOWLEDGE TRANSFER IN THE BORDERLAND

A key competitive advantage of a contemporary economy, knowledge, is distributed unevenly, tending to concentrate in cities and urban agglomerations. A border position translates into distinctive features of regional innovative development. In a favourable institutional context, proximity to a border strengthens transboundary cooperation and interaction between neighbouring regions. Although frequent social contacts across borders are well documented in the literature, the effect that the border has on intensive knowledge transfer is yet to be investigated. This article analyses models of knowledge integration taking place between Russia’s northwestern regions and the countries that their border. The study covers six territories of the Northwestern federal district (the Republic of Karelia, St Petersburg, and the Kaliningrad, Leningrad, Murmansk, and Pskov regions); five regions of the Central federal district (Belgorod, Bryansk, Voronezh, Kursk, and Smolensk); and one region of the Southern federal district (Rostov). The methodology of the study consists of using information from the Scopus abstract and citation database to assess the intensity of research cooperation. The findings suggest that the degree of involvement in transboundary research cooperation varies widely across Russia’s border regions.


Introduction
Literature offers two opposite views on the innovative development of border areas and their contribution to the national innovation system (NIS).
On the other hand, border regions are natural contact zones that interact with elements of the spatial socioeconomic and innovation systems of neighbouring states. These areas can be considered as strategic development corridors. An drey Klemeshev and Gennady Fedorov [22][23] emphasise that border areas have a major role in developing bilateral (Russia-EU) international innovation flows. The contact function of the state border manifests itself in the intensity of movement of material goods (commodities, services, capital), people, and intel lectual capital (knowledge, culture, competencies). This function is also essen tial for international ties [24]. Neighbouring regions of two countries are usually quite close in sociocultural and institutional terms. This kinship creates a favour able environment for stable crosscountry cooperation and enhances the ability to embrace new knowledge and disseminate it nationwide, as well as to adapt it to local conditions) [25]. Stronger integration between neighbouring regions of two countries helps to accumulate a critical mass of participants in innovations to ensure competitiveness at national (NIS) and global level. Researchers have described many cases of successful transboundary regionalisation of innovative systems of border areas in Europe. These examples include transboundary clus ters: Øresund between Denmark and Sweden [26][27][28][29], the Alsace BioValley be tween France, Germany, and Sweden [30], and others. According to [31][32], transboundary regionalisation is a strategic priority in developing the region al innovation systems (RIS) of border areas: regionalisation makes it possible to change dramatically the existing development trajectory and shift the balance of the centreperiphery model of innovation production.
The duality of research findings precludes a conclusive answer to the question about the role and place of border regions in developing the NIS. In particular, researchers have stressed the need for devising a new approach to evaluating the innovative performance of border areas. This new technique should take into account indicators other than the critical mass, network density, and their likes used in studying central regions [33]. In this article, we seek to provide a compre hensive evaluation of the intensity of transboundary research cooperation involv ing Russia's western borderlands. Concentrating on one RIS component makes it possible to measure the engagement of local actors in crosscountry networking, evaluate the quality of intellectual collaborations, understand the localisation of interacting parties, and identify the role of the geographical factor in the distri bution of knowledgeintensive activities. Special attention is paid to the potential for strengthening transboundary cooperation and integration. To this end, we an alyse the unilateral use of research findings by international academics. The key hypothesis of our study is that most of the potential of the RIS research compo nent remains untapped in the states bordering Russia to the east.

Literature review
The internationalisation of research is a result of the growing trend for inno vation processes to become more complicated and for R&D to accelerate. To an extent, these processes are a result of the shrinking product lifecycle and com mercialisation period. In an open market when information on the potential of possible counterparties is widely available, the role of effective management of innovation processes aimed at developing and introducing complementary 'value propositions' is increasing [34][35][36]. State-of-the-art hi-tech infrastruc ture including laboratory and experimental facilities, top specialist and the ability to recruit and train them, as well as business and intellectual excellence may con tribute to transboundary research cooperation. How much international partners are committed to forging partnerships with regional actors depends directly on the expected synergistic effect, the presence of which indicates a significant in crease in efficiency. Despite geographical proximity, the research components of a border area RIS can be so unlike that the complementarity of key strategic development areas is impossible [37][38]. At the same time, absolute compat ibility, which means identical infrastructure as well as similar competences and challenges, makes cooperation less attractive an option. Research cooperation between border region agents requires that they should be at a similar level of development as regards the range of complementary competencies in question (particularly, in terms of disciplinary microspecialisations), explore similar problems, and share a common research paradigm. According to [39], similar ap proaches and technology ensure the transfer of knowledge between collaborating parties. Wesley M. Cohen and Daniel A. Levinthal [40] emphasise that the suc cess of cooperation is largely determined by the capacity of the parties to absorb newly acquired information. Key to their concept is prior related knowledge, which is transferred through direct personal contacts. It includes experience, skills, abilities, know-how, and processes. Here, confidence in results is crucial for a decision to cooperate. A major consideration is thus the quality of current findings and their prospects for implementation in joint projects.
In the research community, Quality is closely linked to the elusive category of academic reputation. 1 At the same time, an additional and commonly recognised quality criterion is the status of research periodicals where a university's em ployees publish their findings. The Web of Science international citation database uses the impact factor indicator, the Scopus counterpart of which is CiteScore 2 . The literature has confirmed that findings from international collaborations are associated with higher quality than those from national ones are [41][42][43].
As the geography of cooperation network expands, the cost of maintain ing permanent contacts increases. For instance, André Torre [44][45] writes that information and communications technology meet the need for person al communication only partially and cannot replace it completely. Temporal geographical proximity, which is achieved through working meetings, round tables, conferences, etc., does not ensure the level of engagement necessary for innovative collaborations, particularly, the generation of new basic knowl edge. Only quality results that significantly boost average values ensure return on investment in a cooperation network (finances, time, intellectual efforts, labour, etc.).
The degree of engagement of a wide range of stakeholders on either side of the border in the joint process of knowledge generation describes the develop ment of network ties. The stability of a transboundary RIS (TRIS) is ensured by a pool of various interacting parties committed to longterm cooperation with a prospect of delegating part of equally significant functions to international part ners, including laboratory tests, design and engineering, software development, etc. Accumulating a critical mass of participants in a transboundary cooperation network facilitates devising a coordinated and prospectively common develop ment strategy. It covers, as a rule, investment policy, promising projects, techni cal and operational harmonisation, and qualification requirements (KPI, working conditions, etc.). The proportion of researchers involved in networking reflects the weight given to the area of cooperation in question and readiness to devel op it. An important factor is a favourable institutional environment that can help to streamline transboundary contacts by simplifying the visa regime, modern ising the road network and bordercrossing infrastructure, enhancing passenger communications, etc. Since trust is a sine qua non of R&D collaborations [46][47], socio-cultural projects and joint non-profit organisations have a significant role in the emergence of a TRIS. A major obstacle, however, is the institutional context, which is largely a product of the geopolitical situation. A good environ ment for bilateral contact between regions of two countries translates into the natural comfort of cooperation.
Geographical proximity facilitates frequent social contacts between the mem bers of the research subsystem of a TRIS. Empirical findings suggest that the density of intranetwork contacts depends on geographical remoteness [48]. Localisation and later clustering create a favourable background for mutual edu cation and knowledge spillover supported by informal personal communication [49][50][51][52][53]. Geographical closeness contributes enormously to stronger informal ties [52; 54], trust, recognition of belonging [50; 51; 52], easier information ex change and access to various types of knowledge [55], as well as the solidarity of likeminded individuals and a common identity [53].

Methodology
Methodologically, this study draws on approaches used in contemporary sci entometrics and analyses a vast array of bibliometric data. This way, it becomes possible to get an idea of the dynamics of transboundary research cooperation. The source of the bibliometric data used in our analysis is Socpus -the largest international citation database, which covers findings distributed by over 5000 publishing houses worldwide, including Elsevier, SpringerNature, Wiley, Taylor & Francis, Sage, and others. We analyse six-year-data (2013-2018). The list of indicators includes the number of publications with international coauthors, the total number of authoring teams, and the citation rate (with the field-weighted citation impact, FWCI, taken into account). The latter measure helps to compare the number of publications across different fields of knowledge.
We searched Scopus for research publications, using the following advanced search queries (those below are for the Kaliningrad region (Russia) -Poland pair): "

AFFILCOUNTRY (Russia*) AND AFFILCITY ("Kaliningrad") OR AFFIL-CITY ("Bagrationovsk") OR AFFILCITY ("Guryevsk") OR AFFILCITY ("Gusev") OR AFFILCITY ("Zelenogradsk") OR (AF-ID ("Immanuel Kant Baltic Federal University" 60031254) OR AF-ID ("Kaliningrad State Technical University" 60018744)) AND (LIMIT-TO (PUBYEAR, 2018) OR LIMIT-TO (PUB-YEAR, 2017) OR LIMIT-TO (PUBYEAR, 2016) OR LIMIT-TO (PUBYEAR, 2015) OR LIMIT-TO (PUBYEAR, 2014) OR LIMIT-TO (PUBYEAR, 2013)) AND (LIM-IT-TO (AFFILCOUNTRY, "Poland"))
The search query for each Russian regions included all publishing cities and towns and the names of all research and educational institutions. The result ant publication pool was exported into the SciVal analytics tool for a detailed analysis of international partner organisations. On the list of analysed measures are the organisation types (an institute of an academy of sciences, a university, a commercial organisation, other), as well as the region and the city/town where this organisation was located.
Data processing consisted of three key stages. Stage 1: using the information on the current network of research cooperation, a pattern of current interregional cooperation was established and the intensity of ties in the border area identified. To this end, network relations were built be tween cities. The intensity of these relations was characterised by the volume of coauthored publications and the total number of authoring teams.
Stage 2: based on an analysis of works citing the findings of Russian research er, potential cooperation channels were identified. At this stage, the analysis cov ered all research centres that did not carry out joint research over the reporting period but benefitted from the intellectual results of Russian authors.
Stage 3: a transboundary research cooperation index was calculated using four subindices: Subindex 1: engagement. This index comprised the following measures: X1, the ratio of joint publication authors from a neighbouring country to the coun try's organisations listed in affiliation; X2, joint publications with authors from a Russian border region as a proportion of all publications of the neighbouring country's research centres that have at least one joint publication with a research centre in the Russian region.
Subindex 2: commitment. This index consisted of the following measures: X3, the ratio of citations of a publication from a Russian region by researchers from a neighbouring country to all research publications in the neighbouring country; X4, the ratio of researchers from a neighbouring country, citing publications from the Russian region, to all researchers from the citing organisations in the neigh bouring country.
Subindex 3: quality. This index covers the following measures: X5, citations to publications coauthored by researchers from a Russian region and the neigh bouring country (citations per paper); X6, the average FWCI of publications coauthored by researchers from a Russian region and the neighbouring country.
Subindex 4: localisation. This index is based on the following measures: X7, the ratio of cities where the research centres of affiliation of neighbouring coun try's coauthors are situated to all borderland cities that are home to universities; X8, the ratio of organisations of affiliation of a neighbouring country's coauthors to all research centres in that country.
The measures were normalised by linear scaling to the range [0; 1], where 0 is the minimum and 1 the maximum attribute value. The initial data normalisation formula for the measures of positive attributes is as follows: where Z ij is the normalised value of the j th measure for the i th region; a �� is the values of the j th measure of the i th region; a � ��� is the maximum value of the j th measure; a � ��� the minimum value of the j th measure.
The subindices and the integrated index are calculated using the arithmetic mean: where Z �� ��� is the value of the integrated index; Z ij is the normalised value of the j th measure for the i th region; n is the total number of measures (n=2 for the subindices and n=4 for the integrated index).
The validation of the measures considered within the above subindices is given in some economic-geographical studies, including those by Russian authors [56][57][58]. Methodological limitations include the emergence of extreme values that were excluded from calculations and the cases when no publications were cited and the FWCI was zero (such observations were not analysed further).

Results
The total number of Russian works indexed in the Scopus international data base in 2013-2018 was 447,818, which places the country 13 th worldwide. The contribution of the border regions under study to the total number of publica tions in Russia is insignificant (Table 1).   . 1, 2). Based on the data on the current and potential structure of research ties, the transboundary research cooperation index was calculated ( Table 2). The com putation did not include regions that have fewer than ten publications with the neighbouring states: the Bryansk region-Ukraine, the Pskov region-Belarus, the Pskov region-Latvia, the Pskov region-Estonia.

Discussion and conclusions
In this study, we examined nineteen geographical areas of transboundary research cooperation between twelve Russian border regions and eight neigh bouring countries in 2013-2018. We evaluated interactions between each of the Russian regions and the neighbouring country (ies) in general, carried out a comparative analysis, and identified the most active areas of research collab orations.
Cooperation with Estonia involves three Russian regions: St Petersburg, the Leningrad region (with which the country has forged stable transboundary ties), and the Pskov region (it had only nine joint publications over the studied peri od). Although the number of Estonian research centres whose experts publish their findings in Scopus-indexed periodicals (seventy-four) is rather small, most of them cooperate with their counterparts from Russian borderlands, primarily, St Petersburg (10.8%). The geographical scope of the Leningrad region-Esto nia partner network is smaller. The level of engagement, however, is onethird higher than that in St Petersburg, being the highest among all the research coop eration areas considered in this article. According to the quality and commitment subindices, research cooperation in the Russian-Estonian borderlands is valued by both parties, whereas its outcomes meet high standards.
The above is explained by stronger contacts in biomedicine, which is a tra ditional research area for Estonia [60] and a current strategic priority for both countries. Key research partners from Estonia are the National Institution of Chemical Physics and Biophysics (Tallinn) and the University of Tartu. The latter is home to the Estonian Biocentre, which was established in the Soviet period. Today, Estonia is one of the few countries in the world that has a suc cessfully functioning genome biobank. In 2015, with financial support from the Russian Science Foundation, St Petersburg State University (SPSU) launched the project to create the first Russian biobank -a dedicated crystorage for bi ological materials and a clinical laboratory for biomedical studies in health and longevity. Alongside the key Estonian Universities, Tallinn University, Tallinn University of Technology, and the Estonian University of Life Sciences, the partnership involves the National Institute for Health Development, which con ducts population studies in healthcare. The St Petersburg agglomeration is a strong research centre that has attracted considerable resources from across the country, including those of Russia's leading universities topping international rankings ( Cross-border cooperation programmes, including those co-financed by the EU (Russia-SouthEast Finland, Kolarctic, and Karelia), and the Interreg Baltic Sea Region programme ensure the stability of transboundary contacts. Many of these projects support research and knowledgeintensive innovations in environ mental protection and ecology. The practical focus of transboundary coopera tion is a clear advantage of Russian-Finnish contacts, which gave rise to trans boundary clusters of cleantech, energy, and timber companies, thus contributing to the stability of transboundary cooperation. Among the cities involved in active cooperation are Helsinki (the University of Helsinki, the Finnish Meteorologi cal Institute, Natural Resource Institute Finland, etc.), Kuopio (the University of Eastern Finland), and the border of Lappeenranta, the administrative centre of South Karelia (Lappeenranta University of Technology).
Norway is a transboundary partner of the Murmansk region. The transbound ary research cooperation index for the two areas is 0.273, which is above that for Murmansk-Finland collaborations (0.217). Although the difference in the in dex values is significant, the disparity between some important measures is even greater. The difference is 1.88fold when it comes to the number of coauthors from the neighbouring country, 1.80fold in the case collaborating organisations, 1.89fold in that of the number of joint publications, and 2.26fold in that of the number of citations. At the same time, Finland outperforms Norway in some as pects of research interactions with Russian regions: the FWCI (2.39 against 1.68) and the number of networking cities (a threefold difference).
Common research areas are a key factor for cooperation. In this case, these are marine resources, ecology, and Arctic studies. A priority for both countries, projects in these areas receive considerable support. The city of Tromsø, which is located relatively close to Murmansk, is one of the key partners in transboundary cooperation that brings together the University of Tromsø -The Arctic Uni versity of Norway and Norwegian Polar University. The latter organises Arc tic expeditions and conducts research at the NyÅlesund station on the island of Spitsbergen. Some of the expeditions are joint initiatives supported by the Arctic Council, the Nordic Council of Ministers, the Northern Dimension, and other in stitutions. Other important partners are the Institute of Marine Research (Ber gen), Norwegian Institute for Water Research (Oslo), Centre for International Climate and Environmental Research (Oslo), the Geological Survey of Norway, SINTEF (an independent research organisation that conducts contract research and development projects), and Equinor ASA (Norwegian international energy company). These organisations are located at a significant distance from Russian border regions.
Poland is a partner of the Kaliningrad region, Russia's Baltic exclave. This cooperation area has the lowest engagement subindex. Twenty-five organisations are involved in cooperation, yet only threefour researchers from each contribute to crossborder collaborations. Moreover, the proportion of joint studies does not exceed 0.04% of all studies. To a degree, this is explained by the intensive pub lication activity of Polish research centres collaborating with the Kaliningrad re gion (over 132. Active interactions are supported by transboundary mobility. An important factor in the latter was the local border traffic regime between Poland and the Kaliningrad region. Contacts in the field of maritime economy are developed by Kaliningrad research centres specialising in the area: the Shirshov Institute of Oceanology of the Russian Academy of Sciences, the Atlantic branch of the Russian Research Institute for Maritime Economy and Oceanography, the Baltic Fish Fleet State Academy, and the Museum of the World Ocean. Research teams of the Institute of Regional Studies at the Immanuel Kant Baltic Federal Uni versity (IKBFU) also contribute to the process. Biomedical collaborations are supported by the IKBFU's laboratories, including those located at the Fabrika science park.
Only four Lithuanian research centres have formed partnerships with the Ka liningrad region: Klaipėda University (13 joint publications), Vilnius University (9), Kaunas University of Technology (2), and Vytautas Magnus University (1). The Lithuanian cooperation area has higher engagement (0.047) and commit ment (0.073) indices than the Polish one. Firstly, interactions involve at least eight people per organisation, which is twice the Polish level. Secondly, find ings obtained in Lithuanian-Kaliningrad collaborations account for 0.33% of all Lithuanian publications, which is 86.8% above the Polish proportion. Thirdly, the rate of citations of Kaliningrad authors by Lithuanian colleagues is higher than by the Polish ones (0.44% and 0.05%). Alongside the research centres in volved in cooperation, citing organisations include the Lithuanian Energy In stitute of the Lithuanian Academy of Sciences (Kaunas) and Vilnius academic organisations: Mykolas Romeris University, Vilnius Gediminas Technical Uni versity, National Centre of Physical and Technological Sciences. All of these research centres are potential partners. The quality subindex is close to average Collaborations with Ukraine are pursued by a considerable number of Rus sian borderland organisations from the Belgorod, Bryansk, Voronezh, Kursk, and Rostov regions (table 2). At the same time, these regions score among the lowest on the transboundary research cooperation index for 2013-2018 (ta ble 2). The Bryansk region was excluded from the calculation because it had not reached the threshold value of ten joint publications. The levels of com mitment and engagement are also rather low ( Table 2). The Russian-Ukrain ian crossborder cooperation involves only 6% of the Ukrainian research cen tres visible to Scopus. Most of the Ukrainian organisations collaborate with the Rostov region (26) and the fewest with the Kursk region (11). At the same time, the ties of the Kursk region cover all the university cities of Ukrainian border lands, whereas the Rostov region collaborates only with 38% of those (three out of eight) and the Belgorod region with 40%. The Voronezh region is the only border area in Russia that does not cooperate with the border research centres of a neighbouring country.
Research cooperation with the Kursk region focuses on materials science and physics. Key Ukrainian partners are the cities of Sumy (Sumy State Uni versity) and Kharkiv (Kharkiv Institute of Technolgy, the Karazin National University of Kharkiv, the Verkin Institute of Low-Temperature Physics, and the Institute of Radio Astronomy of the National Academy of Sciences of Ukraine [NASU]).
The Belgorod region also seeks cooperation with the border cities of Kharkiv (Kharkiv Institute for Physics and Technology, the Karazin Nation al University of Kharkiv, Kharkiv Institute of Technology, Kharkiv National University of Radio Electronics, the Institute of Cryobiology and Cryomed icine of the NASU, Usikov Institute for Radiophysics and Electronics of the NASU), and Sumy (Sumy State University). Alongside materials science and electrophysics (colloid chemistry, electrowelding, and solidstate physics), the Belgorod region collaborates with Ukrainian research centres in medical and biological fields (cell biology, cryobiology, biochemistry). The number of Bel gorodUkraine publications is one of the highest in the studied regions (298) This study revealed several patterns in transboundary research cooperation: -intense research cooperation was observed in regions that boasted an equally high level of research and had considerable commitment to international partnerships; -complementary competencies and knowledge bases within a common re search area was a major factor for forming transboundary ties; -a strong impetus for developing research cooperation in the borderlands was programmes aimed to support transboundary projects and joint studies; -the population pattern, economic clustering, and transport connectivity made frequent personal contacts possible and thus contributed to forming re search ties; -a developed transboundary institutional environment, a favourable geopo litical situation, and cultural proximity were important factors in strengthening research cooperation.