Strategic Research Partnerships (SRPs), in the form of technology-based joint ventures, strategic alliances and multi-partner R&D projects, are an important feature in the generation and diffusion of technology and, by extension, industrial development. They are an important feature of the research environment and industry in most industrialized and industrializing nations. This paper examines some explanations for their formation, and some issues related to the measurement of the extent and the outcomes of SRPs.
The paper focuses particularly on East Asian and European experiences with SRPs. It argues that while there are some common motivations underlying the formation of SRPs internationally, there are broad national differences in the role they play. There is also varying capacity in government and research organizations to quantify and measure the contribution of SRPs.
As SRPs are created for a variety of purposes, and assume a range of different forms (differences that are accentuated when international comparisons are factored in) there are immense methodological problems in measuring their extent and contribution. The paper illustrates the wide range of indicators that can be used to examine specific features of SRPs. The types of analyses discussed include: science indicators and bibliometrics, international and national surveys, specific databases on alliances, network analysis and various forms of case study. There are shortcomings and deficiencies in these measures, and the value in them for policy-makers depends upon their imaginative combination in ways that address specific concerns.
In the following discussion it must be appreciated that there are major shortcomings in the collection of the various forms of information in most East Asian nations.
SRPs are understood here to essentially involve shared commitment of resources and risk by a number of partners to agreed complementary research aims. SRPs can occur 'vertically' throughout a value chain, from the provision of raw materials, through the design, production and assembly of parts, components and systems, to their distribution and servicing. 'Horizontal' SRPs, on the other hand, occur between partners at the same level in the value chain.
SRPs between firms can take a variety of forms. They may be a joint venture, formed by two or more partners as a separate company with shared equity investments. They can be a partnership or 'strategic alliance' linking firms on the basis of continuing commitment to shared business or technological objectives without equity sharing. They may take the form of R&D contracts or technology exchange agreements whereby firms' shared objectives involve the interchange of research findings or technological know-how. Universities and public research laboratories are often partners in such R&D contracts. SRPs may take the form of 'innovation networks', combinations of firms and research organizations that share research agendas.
There is a wide range of explanations for why firms and research institutes collaborate in their research activities. There are economic explanations (cost reduction and efficiencies), and those that consider the strategic competitive relationships between firms (standards creation, competitor exclusion or locking-in key players). Some address technological issues (the way the importance and role of SRPs change with the technology life-cycle), while other explanations are less instrumental and focus on qualitative issues such as organizational learning (see Dodgson, 1993a, for a discussion of these various approaches). Hagedoorn et al (2000) separate three traditions in theory that explain research partnerships: transaction costs, strategic management, and industrial organization. There is diversity within each tradition, so, for example, within the strategic management field it is possible to consider approaches that emphasize competitive forces, strategic networks, resource-based theories, dynamic capabilities and strategic options (Hagedoorn et al, 2000).
A firm- or institution-level perspective is required to analyze the motives for forming SRPs. In what follows a brief analysis is provided on why these partnerships are created from the perspective of the firm, where SRPs are seen as a means of improving technological competencies and learning about new markets, management practices, and strategies.
Although SRPs occur in many different forms, and may reflect different motives, a number of generalizable assumptions underpin them. First, is the belief that SRPs can lead to positive sum gains in internal activities. That is, partners together can obtain mutual benefits that they could not achieve independently. Such benefits may include the following.
Increased scale and scope of activities. The outcomes of SRPs may be applicable to all partners' markets, and thus may expand an individual firm's customer bases (increased scale). Synergies between firms' different technological competencies may produce better, more widely applicable products (increased scope). Increasing the scale of resources dedicated to research programs can also raise entry barriers to other firms.
Shared costs and risk. SRPs can share the often very high costs, and therefore risk, of research (although they can also, of course, share future income streams from any subsequent innovations).
Improved ability to deal with complexity. Closer strategic and technological integration between firms and research institutes is a means for dealing with the complexity of multiple sources and forms of technology. It allows, for example, the better transfer of tacit knowledge.
A second assumption regarding SRPs concerns the way they assist with environmental uncertainty. Increasingly sophisticated and demanding customers, growing competition in and globalization of markets, and rapidly changing and disruptive technologies place pressures on firms to exist with, and attempt to control, these uncertainties. This is believed to be more easily achieved through partnership than in isolation. Strategy is a means of dealing with uncertainty, and SRPs allow firms to observe and transfer useful lessons about strategy from partners. An important contextual factor affecting the level of uncertainty is the increasing level of scientific and technological integration occurring in various forms. Kodama (1995) discusses the increasing prevalence of 'technological fusion'. Thus 'mechatronics' involved the fusion of mechanical technology with electrical and material technologies, and 'optoelectronics' involves the fusion of glass technology with cable and electronic device technologies.
A third set of assumptions underlying SRPs concerns their flexibility and efficiencies compared to the alternatives. For example, SRPs may be an alternative to direct foreign investment, mergers, and acquisitions which are much less easily amended once entered into. As a governance structure, SRPs possess advantages over the alternatives of arms' length transactions and vertical integration. They can allow firms to keep a watching brief on external technological developments without having to invest heavily. Large firm/small firm interaction can be facilitated such that the resource advantages of the former are linked with the behavioral or creative advantages of the latter whilst maintaining their independence (Dodgson and Rothwell, 1994). Large drug companies, for example, may partner with a small biotechnology firm as a means of developing their options so that they could invest more heavily once the technology is better proven and better understood. The larger firm will have gained the opportunity to learn about the technology during the SRP.
While information and communications technologies have facilitated increased and more effective SRPs, much technological knowledge is not only tacit, but firm-specific (Pavitt, 1988). It is, therefore, difficult to transfer easily or quickly. SRPs potentially provide a mechanism whereby close linkages among different organizations enable the development of sympathetic systems, procedures, and vocabulary which may encourage the effective transfer of technology. It may also allow partners to 'unbundle' discrete technological assets for transfer (Mowery, 1988). Finally, SRPs may address the difficulty of valuing technological knowledge by providing a means of exchange that does not necessarily rely on price.
Potentially, therefore, there may be numerous advantages to be achieved through SRPs if these assumptions hold. These benefits are not only economic, but also behavioral: firms can learn about new markets, technologies and management practices through SRPs.
There are also potentially adverse aspects of such partnerships. SRPs can be anti-competitive, by excluding certain firms, or raising entry barriers, or operating in the form of cartels which anti-trust legislation prevented in the past. Also there may be strategic dangers for firms which overly rely on externally sourced rather than internally generated technology. Without internal technological competencies there can be no 'receptors' for external technology, nor capacity for building the technological competencies which provide the basis for firms' technology strategies (and which provide the basis for attracting potential partners). In addition to the positive benefits of innovation networks they also can have negative consequences. The network model of innovation may limit participating firms' access to 'complementary assets' (Hobday, 1994) and hence their ability to achieve full commercial returns to innovative activity.
National variations in the extent to which SRPs occur are influenced by differences in the role of governments, industrial structures, business systems and research infrastructure within national innovation systems.
There is a great deal of similarity amongst industrialized and industrializing nations in policies towards the encouragement of SRPs. The encouragement of SRPs is a key policy focus of the European Commission, as seen in policies such as ESPRIT (a collaborative program with an IT focus) and the five Framework Programs conducted between 1984-2002 (funded collaborative research in a range of industries). In the United States, SEMATECH provides an example of government-sponsored research partnership and, in Canada, the IRAP scheme encourages collaborative research between firms and universities. A wide range of SRP-promoting policies are also found in Japan, ranging from large-scale, high technology schemes that began with the fifth Generation Computer Project to local support schemes through over 150 Regional Technology Centres. Taiwan's Industrial Technology Research Institute (ITRI) has played a central role in encouraging technological development and diffusion through collaborative projects. SRPs are supported by policy-makers internationally as a means of building the inter-firm and research networks that are essential elements of an innovative economy (Dodgson and Bessant, 1996).
There are, of course, major national differences in the policy objectives and constraints regarding SRPs in the extent to which they are driven by scientific, technological and competitive objectives. Korea's policies of supporting national champion firms through a variety of forms of SRP would be anathema to most industrial policy-makers in the West. Within the European Union, there are broad policy differences towards SRPs in accordance with different science and technology policy systems (Rothwell and Dodgson, 1990) and policy-making processes. Nevertheless, the broad policy support for SRPs, albeit with widely differing degrees of support, focus and intent, is a common motivating factor for firms and research institutes to collaborate.
The extent to which firms source research externally, both vertically and horizontally, is affected by particular national industrial structures. A commonly cited reason for the high levels of external integration in Japanese industry, for example, is the structure of industry itself; in particular the role of the Keiretsu and the strong vertical relationships found in Japan down the supply chain. The structure of Taiwanese industry, with its predominance of smaller firms, has encouraged the amount of collaborative research, particularly that centered around ITRI. Similarly, the large number of small and medium-sized firms in German manufacturing industry has encouraged the extensive use of industrial research associations (Rush et al, 1996).
Business systemsthe ways in which firms relate to one another, to their employees, government and to financial systemsvary so significantly that different kinds of capitalism can be described (Dore, 2000). These systems affect the general propensity towards cooperation, and will influence the extent and role of SRPs. Differences between 'Anglo-Saxon', 'Rhine' or Japanese capitalism will be seen in the breadth and depth of SRP activity. This can be a reflection of the differing strategies of firms as they relate to others along a continuum of spot-trading to 'obligational contractual relations' (Sako, 1992). These differences are particularly important when consideration is made of Chinese capitalism (Redding, 1993). In addition to its ubiquity in China, Taiwan and Singapore, Chinese business practices, which are strongly family-based, are dominant in Malaysia, Indonesia and the Philippines. Chinese family businesses have a strong preference for doing business with people that are associated through kinship or geographical origins. This can have the effect of limiting options for SRPs.
The propensity to conduct SRPs is obviously affected by differing research infrastructures and the way these are integrated within national innovation systems. National science and technology capabilities, as determined by levels of R&D expenditure and employees, investment in universities, etc., vary significantly within and between Europe and East Asia. Also important is the level of integration between the different players in the innovation system. The existence of a range of strong, well-established research institutes, experienced at working with industry, such as the Fraunhofer Society institutes in Germany, are important elements of national innovation systems, and encourage SRPs. By contrast, the comparative lack of development of SRPs in China (outside of technology contracts) is caused by the historically almost complete disengagement of research institutes from industry.
Whereas most European nations enjoy access to strong national and pan-national research institutions and firms, many East Asian nations remain impoverished in this regard. Within East Asia there are massive differences in science and technology capabilities, seen particularly clearly in disparities in R&D expenditure and employment (Dodgson, 2000b). Whilst Singapore, Taiwan, Korea and Japan have developing research infrastructures, particularly in some industries, and relatively coherent national innovation systems, other East Asian countries do not possess the capacity to undertake SRPs. With countries, like Indonesia and Thailand, spending around $2 per capita annually on R&D, SRPs are only likely to be a marginal concern for the limited number of science and technology-based organizations and firms. However, whereas the sort of research partnerships found in developed economies based on 'pre-competitive' R&D is likely to be extremely rare in these countries, the more 'diffusion-orientated' partnerships are of central importance to the development of the national technology base (Dodgson, 2000b).
Policies towards, and funding for, research institutions dynamically affects the extent to which SRPs occur. In Australia, for example, the combination of budgetary constraints affecting universities, legislation mandating that the nation's largest government research organization, the CSIRO, obtain 30 per cent of its budget from industry, and initiatives such as the Cooperative Research Centers that encourage business/research links has led to historically high levels of SRPs. Such changes are common internationally, but have perhaps been seen most radically in China where, as a result of changing policies, a massive change is occurring within research institutes, and in the productive sector, in the extent and form of SRPs.
Changing policies towards research institutions not only affect the extent of SRPs, but also their intent. Seen particularly in countries like Taiwan and Korea, the national research institutions have had to adapt and change in their research activities as some industries have moved from positions of technological following to technological leadership (Kim, 1997; Dodgson, 2000b).
Measuring the scale and importance of SRPs is notoriously difficult. Data on their extent and outcomes is often piecemeal and occasionally contradictory. Furthermore, whereas the bulk of evidence suggests an increasing role for SRPs in industry, the majority of studies of their outcomes point to the considerable difficulties in gaining mutually satisfactory outcomes amongst partners in collaborative research projects (Dodgson, 1993a). These difficulties are often more apparent in horizontal partnerships as these may more often lead to disputes over ownership of their outcomes, such as intellectual property rights, or to direct competition between partners. Many of the problems of measurement relate to the ways in which the objectives of SRPs can change over time, reflecting the learning that has occurred in the partnership (Dodgson, 1993b).
However, a range of different measures does exist. As we shall see, they have varying utility in measuring the extent, conduct and outcomes of SRPs.
Mapping techniques based on patent and bibliometric data are being used to analyze the structure and dynamic development of scientific and technological developments, including the growing inter-relationship or fusion of areas of science and technology. These indicators not only measure the 'context' or environment in which SRPs occur, but also directly record the SRPs of individual companies and research institutes.
Using information about the content of patents, or scientific publications, enables mapping and visualization of the cognitive structure of specific scientific or technological areas. Most recent methodological developments and highly sophisticated applications of mapping techniques can be found in Noyons (1994).
Many areas of technology are characterized by a close relationship to science, which, according to recent science and technology foresight studies, will even increase in the years to come. Analyzing those relationships is important for firms as it significantly influences the generation of new technologies. The degree of the science dependence of a certain area of technology, or the degree to which an area of technology is science based, can be measured using an indicator calculated by using the citations to scientific publications given in the official search reports of patents.
Sufficiently large samples are required to make the results meaningful (Schmoch, 1997), and that is why, generally, the indicator is used for whole areas of technology. However, the company, CHI Research, also applies this indicator at the firm level to measure the science linkage of individual firms. CHI claims that this indicator allows the identification of the high tech players in certain fields of technology and that the indicator 'has been found to be predictive of a company's stock market performance' (CHI Research, 1999). Multiplying the science linkage with the total number of patents of a firm leads to the science strength of this company, which is seen as indicating the 'total amount of a company's science linkage activity' (CHI Research, 1999).
Having a high share of patents applied for by scientific institutions is also seen as giving an indication about close relationships between scientific and industrial activities. According to Schmoch (1997) due to the high costs of patent applications for scientific institutions it makes only sense to apply for a patent if there is an interest in the further commercial exploitation of the invention, which means that collaboration with industrial partners already exist or are intended.
Scientific publications by industrial enterprises might similarly be used as an indicator for existing relationships between science and technology. Those publications are seen as signaling scientific competence from the industrial enterprise and an interest in getting involved in scientific communication in the specific area. Direct collaboration between scientific and industrial institutions can be measured by co-publication of either publications or patents.
Cross-disciplinarity in research activities, which is one of the reasons for SRP formation, is being measured using information about the institutional affiliation of the authors in multi-authored scientific publications. Bourke and Butler (1995), for instance, showed that cross-disciplinary behaviour increased between the early eighties and the early nineties. Hinze (1999) showed that in the area of auto-immune disease research there are differences between the proportion of cross-disciplinary research depending on whether research is carried out within or across the borders of an institution within a country.
Science is increasingly internationalized, with an increasing proportion of academic publications being derived from international collaborations, particularly in basic research (Bourke and Butler, 1995). Numbers of databases measure the extent to which scientific publications are produced with foreign co-authors and patents are registered with foreign co-inventors (OECD, 1999b). There are significant international variations in these data. Around 50 per cent of scientific publications in Hungary, Portugal and Switzerland are undertaken with a foreign co-author, compared with the OECD average of 27 per cent, EU average of 18 per cent, and Japan's 14 per cent. 83 per cent of Turkey's patents are registered with foreign co-inventors, compared the OECD average of 9 per cent.
There are some dangers in using bibliometric techniques in measuring SRPs. Multiple-authorship is widely used as an indicator to measure research collaboration. The underlying assumption is that the authors involved carried out the research leading to the paper in collaboration. Results of these bibliometric analyses should, however, be interpreted with caution. Katz and Martin argue that while '...the assessment of collaboration using co-authorship is by no means perfect, it nevertheless has certain advantages' (Katz and Martin 1997: 3). According to their argument multiple-authorship should only be used as a 'partial indicator' for analyzing research collaboration because only those activities that eventually lead to a jointly authored scientific publication are taken into account and included in the investigation. Not all collaborations, however, result in publications and, conversely, a joint paper does not always mean that the results presented in the paper are based on research collaboration. At the more applied end of the R&D process collaboration may be measured using patent data in a similar manner. Basically, the same shortcomings already mentioned for publication data apply.
A range of international and national surveys are conducted that contain data on SRPs, or SRP-like activities (although the paucity of research activity, and the incapacity to record that there is, in many East Asian nations must be recalled). The Annual World Competitiveness Yearbook, for example, surveys its respondents about whether technology transfer between companies and universities is sufficient, and whether technological cooperation between firms is common or lacking. table 1 shows the results from this survey for Asian and European countries. It shows that countries like Singapore and Taiwan are assessed to do comparatively well in the effort to which firms collaborate with other firms and universities.
Statistics on the extent of various forms of SRP can be derived from compliance with government reporting requirements in a number of areas. Participation in the various types of SRP promoted by the European Commission provides another obvious source of data. China, in its Science and Technology Indicators, records the number and value of domestic technology development contracts between buyers and sellers of technology. These increased from 34,174 contracts worth 7,000 million Yuan in 1993 to 41,019 contracts worth 11,600 Yuan in 1997 (MOST, 1998). Generally, however, government-collected statistics in East Asia do not attempt to measure the extent of SRP activity.
Firm level surveys, like the EC Community Innovation Survey (CIS), also provide information about R&D cooperation but at a rather general level. The data from CIS are derived from a survey of 33,700 enterprises, in 12 EU countries, produced by Eurostat, the Statistical Office of the European Communities. Preliminary results of the second CIS, conducted in 1996, are currently available. Eurostat advises that as the results are preliminary and that all countries are not included, a certain caution should be exercised in drawing too extensive conclusions from the comparisons between countries. The survey is mostly directed towards assessing the percentage of innovative companies (determined by reliance on new products) in the various countries. It shows that firms rarely innovate alone.
Similarly the PACE and Yale surveys focus on the sources of technical knowledge of firms, and shows the importance for firms of research links with research institutes (Klevorick et al, 1996).
Some European studies have analyzed the comparative importance of international collaboration (DeBresson, 1997, Report on the Focus Group on Innovative Firm Networks, 1998, quoted in OECD, 1998). Tables 2 and 3 show these data.
A number of databases measure the numbers of new international technology alliances announced in the technical press. These tend to cover high-profile, technology creating projects and under-represent more technology diffusion-oriented partnerships and those based outside non English-speaking countries. The best of these databases, the MERIT-CATI database, which includes inter-firm collaborations, shows the increase in the number of new collaborations being formed throughout the 1980s and 1990s (see Figure 1). The majority of these new collaborations occur in new technologies, particularly in IT, and are based in the United States, Japan and Europe. Although there has been an increase in partnerships outside of the Triad, primarily in technologically advanced East Asian nations, these still only account for around 20 per cent of the total recorded number (Duysters and Hagedoorn, 2000; Hagedoorn et al, 2000). The major drawback of these databases is that they rely on information that has to be reported by more or less publicly available sources and thus the proportion of unreported or confidential agreements is unknown. They may also underestimate the extent of alliance activity amongst technologically advanced East Asian and Latin American nations. Other, US based, databases include the CORE and NCRA-RJV databases (Hagedoorn et al, 2000).
An example of a specific database containing information about alliances in a particular technology is provided by Recap (Recombinant Capital), which focuses on alliances in the area of biotechnology. Using Recap it is possible to differentiate between the type and origin of partners involved in an alliance (pharmaceutical firm, biotech firm, research institute, university); the type of activity of the alliance; and the stage of the innovation process, which is rather clearly defined in the area of pharmaceutical research due to special approval requirements. Data can also be analyzed for different sub-fields of biotechnology. Information included in the database is gathered from the US Securities and Exchange files where according to Recap 'over 50% of biotech agreements with major pharmaceutical companies, universities or other biotech companies are filed' due to public filing requirements (Recap 1999). Additional information is added originating from publicly available sources like press releases and the Clinical Trial Progress Database. Information in Recap can be retrieved via the internet (Dodgson and Hinze, 2000).
Networks of one sort or another are powerful mechanisms for communication and the transfer of complicated information and technology flows. Networks can enable the sharing of resources, for example, specialist equipment or R&D projects where the costs and risks of investment to any individual firms would be prohibitive. Definitions of network vary, but here they are considered to be an open system of interconnected firms and institutions with related interests (see Castells, 1996). Networks offer a rich web of channels, many of them informal, and have the advantage of high source credibilityexperiences and ideas arising from within the network are much more likely to be believed and acted upon than those emerging from outside. They are therefore an effective mechanism for encouraging learning, an objective of SRPs. Their formation has been a major innovation policy objective around the world (Dodgson and Bessant, 1996).
The measurement of the extent and outcomes of networking activity as it applies to research is very difficult. For example, while there are data on the number of suppliers a particular firm may have, and firms know how much of their R&D is undertaken externally, there is rarely detailed information on the importance and nature of particular links with suppliers or collaborators. There is, however, some good European research into networks which shows not only the extent of networking (see Tables 2 and 3), but also begin to delineate the various types of network in the extent to which they involve equipment suppliers, users, competitors, component suppliers and government laboratories and universities (DeBresson et al, 1997, quoted in OECD, 1999a).
The most successful research project in this area is the Danish System of Innovation in a Comparative Perspectivethe DISKO-project (which provided the basis for the data provided in Table 3). The central issues of this project are:
DISKO was launched in April 1997. It surveyed 1,022 firms. The main categories of questions were:
The major findings from the DISKO study firms showed that inter-firm co-operation in product innovation was frequent, including amongst small firms. There are distinct national variations, and variations in the propensity to cooperate depending on firm size (OECD, 1999a).
The questionnaire developed for this project, by the IKE group at the University of Aalborg, is presently being used in a broader project across a number of other nations.
Case studies of individual firms show both the extent of SRPs and, more than any other indicator, their (often changing) focus, and outcomes. Figures 2 and 3, for example, show the reliance of Samsung Electronics on various forms of technological link with US and Japanese firms. Table 4 also shows they way in which the form of partnership changes as the company becomes more technologically self-sufficient. The focus of the partnerships progress from licensing-in to joint R&D projects (Dodgson and Kim, 1997).
Detailed case studies show the management problems of SRPs, with their implications for different forms of management structures and the importance of inter-partner trust (Dodgson, 1993b; Child and Faulkner, 1998).
Case studies of the management of global corporate R&D show the extent of international SRPs, and some of the major problems that need to be managed to achieve satisfactory objectives (Kuemmerle, 1997; Reger, 1997; Meyer-Krahmer and Reger, 1999).2. Regions
Analyses of particular regions can show the extent of localized SRPs and the importance of particular stimuli, such as strong research institutes or high levels of specific localized government expenditure. They can also reveal the widely different forms of the SRPs, so, for example, Saxenian's (1994) comparison of Silicon Valley and Boston's Route 128 contrasts the centrality of large firms in the latter, with the generally small firm model in the former. Although still in its development stage, Malaysia's Multimedia Super Corridor represents a very large-scale initiative to build regional research partnerships facilitated by substantial infrastructural investment.3. Industries
By conducting in-depth case studies on particular industries it is possible to delineate the most important research relationships. This is exemplified in Mathew's and Cho's (2000) analysis of the electronics industry in Figure 4.
This form of analysis has been instrumental in explaining the successful development of the IT industry in Taiwan. The networks created amongst small Taiwanese firms through their research links with research organizations and international firms have played such an important role in the success of the industry, that Mathews and Cho (2000) have described the model as a new form of economic learning.
An exemplary case study of the role and importance of research linkages in the development of a particular technology is provided by Malo and Geuna (2000). By using a variety of techniques, their study of the development of combinatorial chemistry and biology shows the extent of the knowledge spillovers between science and technology in a research network.5. Policies
Large-scale evaluations of SRP-promoting policies can measure both the extent, motivation and outcomes of SRPs. A classic case in this regard is the evaluation of the UK Alvey Programme in Information Technology (Guy and Georghiou, 1991).
This paper has examined the reasons why firms and research institutions create SRPs. It has considered some methods of measuring the extent and outcomes of SRPs, particularly by reference to indicators found in Europe and East Asia. There are a number of significant methodological problems in measuring the extent and contribution of SRPs. The first of these relates to the diversity of forms which SRP's assume, and the ways in which they can change over the technology life-cycle, or in response to organizational learning. The second is allied to the motivation for forming SRPs. These motives may proceed beyond the purely economic (although economic returns can also be difficult to ascertain in, for example, assessing the relative contribution of internal and collaborative efforts), to the dynamic behavioral changes associated with learning, which is very difficult to measure. The third, is the shortage of governmental and independent academic research into SRPs, seen most particularly in East Asia.
Just as there is no unified theory explaining SRPs, there is no one particular source of data that can be relied upon to provide anything near a complete analysis of their extent, conduct and outcomes. A number of forms of measurement have been described, with varying utility for examining different aspects of SRPs. A much more complete reflection of the extent and importance of SRPs can be gained by combining a number of these different indicators for specific analytical purposes.
Science and technology indicators constructed using data on scientific publications and patents mainly cover the context in which SRPs occur, although individual firm and research institute behavior can also be recorded. Specific databases collecting data on alliances or joint ventures provide further information, which may also reveal the character of the specific SRP.
Valuable additional information, on issues such as the conduct and outcomes of SRPs, can be better tackled using indicators constructed based on firm-level survey data and case studies.
Using a larger number of indicators, from patent and bibliometric data, survey data or other specific databases, and case studies allows the construction of a more comprehensive picture of the extent, conduct and importance of SRPs. However, it will not always be necessary or feasible, due to time or financial constraints, to use all these measures. Decisions have to be made about the appropriate selection of the indicators and these will depend on the definition of the major interests of the investigation undertaken. In any case, the advantages and disadvantages or limitations of the selected set of indicators have to be made clear.
The clear advantages of bibliometric and patent data is the fact that they are easy to access, as the data is contained in publicly available databases; and the analysis of a rather long time series is possible. Innovation surveys on the other hand are carried out on a more or less irregular basis. Only recently has a periodical data collection been started in Europe, the CIS, although there appear to be delays in the conduct of the next survey.
Surveys and case studies of SRPs rarely take account of the aggregate information available from the sources described above. A greater concern to adapt questions such that the individual firm can be benchmarked against these databases would be an opportunity not only to improve the integration of the different approaches, but also to cumulatively build more detailed knowledge about the shortcomings of both approaches.
The policy challenge is not only to improve the quality and reliability of these different sources of data, but to develop metrics for their combination in response to particular policy requirements.
 This assertion is based on research visits to, and collection of available data from, Japan, India, Taiwan, Indonesia, China, Malaysia, Singapore, Vietnam, Philippines, Hong Kong, Brunei and Thailand.
 This section is based on Dodgson (2000a).
 This section is based on Dodgson and Hinze, (2000).
 Although in Europe and East Asia there is no equivalent to the formality of data reporting of a particular term of SRP in the US where, for example, 665 research joint ventures had been registered under the 1984 National Cooperative Research Act by 1996 (quoted in OECD 2000).