This paper investigates the paradox of research and development (R&D) that is being increasingly undervalued by firms and nations. There has been recent debate about whether firms and governments should concern themselves with R&D or instead focus on the much-wider process of innovation. There is a continuing R&D productivity crisis; despite all the heavy investment in research in many sectors, it appears to be yielding less valuable or significant new products, services and processes.
This paper investigates the paradox of research and development (R&D) that is being increasingly undervalued by firms and nations. There has been recent debate about whether firms and governments should concern themselves with R&D or instead focus on the much-wider process of innovation. There is a continuing R&D productivity crisis; despite all the heavy investment in research in many sectors, it appears to be yielding less valuable or significant new products, services and processes.
This paper investigates the paradox of research and development (R&D) that is being increasingly undervalued by firms and nations. There has been recent debate about whether firms and governments should concern themselves with R&D or instead focus on the much-wider process of innovation. There is a continuing R&D productivity crisis; despite all the heavy investment in research in many sectors, it appears to be yielding less valuable or significant new products, services and processes.
Jeremy Howells Manchester Institute of Innovation Research, Manchester Business School, University of Manchester, Manchester, UK. jeremy.howells@mbs.ac.uk This paper investigates the paradox of research and development (R&D), that is being increasingly undervalued by rms and nations, and yet continues to grow and prosper in terms of overall size and reach. The analysis outlines key developments that are currently affecting the growth and development of R&D activity and highlights the issues and problems that R&D managers and policymakers may likely face over the next decade. 1. Introduction T his paper seeks to outline the current and prospective challenges facing research and development (R&D). There has been recent de- bate about whether rms and governments should concern themselves with R&D or instead focus on the much-wider process of innovation. Many of the arguments surrounding this are valid, particularly within the context of advanced economies as they become more service-oriented in nature. Moreover, on a more specic company level, there has been criticism that R&D has just not delivered in terms of innovative output and better productivity. There is a continuing R&D productivity crisis; despite all the heavy invest- ment in research in many sectors, R&D appears to be yielding less valuable or signicant new products, services and processes (Section 2.6). The value placed on R&D by companies seems to be at an all-time low and continues to be declining. In a recent survey conducted by IBM (quoted in Radjou, 2006, p. 7), in-house R&D was only rated 8th in terms of importance as a source for business innovation. Yet despite all these recent arguments and criticisms, R&D continues to develop and grow both within the rm and at a wider national and international level. Why this paradox of criticism and low esteem of R&D at a time when R&D worldwide as an activity continues to grow? This paper will seek to unravel such a paradox and the concluding section of this paper will cover this conundrum. In the context of this debate, there is, however, a danger of neglecting the continuing signicance of R&D for the success and growth of rms and economies. Even in terms of its economic size, R&D is a major activity. In 2000, global R&D expenditure totalled some $729 billion (National Science Foundation, 2006, pp. 440) and in 2006 R&D expenditure was approaching the $1 tril- lion 1 mark. Similarly, R&D has many millions of workers employed in the activity worldwide, with the United States alone employing over 1 million R&D workers. 2 R&D is, therefore, a large-scale endeavour, which continues to grow both abso- lutely and relatively worldwide. The signicance of R&D is also reected at the rm and the government level. The largest com- pany R&D spender, Daimler-Chrysler, spent $7.69 billion alone on R&D in 2005. Meanwhile, governments still remain concerned about their national R&D in terms of spending. This is high- lighted in Europe, with the European Commis- sion having a target of 3% R&D expenditure as a percentage of GDP by 2010 (see Commission of the European Communities, 2002, 2003). The rest of this paper is in two main sections. The rst section outlines some key developments R&D Management 38, 3, 2008. r 2008 The Author. Journal compilation r 2008 Blackwell Publishing Ltd, 241 9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main St, Malden, MA, 02148, USA that are currently affecting the growth and devel- opment of R&D activity. The second section is then more prospective in nature and raises the issues and problems that R&D managers and policymakers will either face in the future or those that are currently an issue, but remain, for what- ever reason, neglected or downplayed by industry and/or government. It should be noted that the main focus of the discussion and analysis of R&D activity will be on business R&D, although as will be indicated public and not-for-prot research is closely intertwined with private R&D activity. 2. Changing dynamics of R&D 2.1. Introduction This section will look at a number of major trends and issues that are shaping the growth and development of R&D globally. These are: 1. the increasingly distributed and open nature of networked research and innovation; 2. this is, in turn, linked to the growth of ex- ternally sourced R&D (and a consequent re- lative decline in internally generated R&D) within rms; 3. overcoming barriers towards the increased productivity and effectiveness of R&D; 4. the continued globalisation of R&D, particu- larly in terms of its spread and reach, asso- ciated with R&D offshoring; and 5. the relative shift from manufacturing-centred R&D towards more service-oriented R&D. 2.2. Distributed and open R&D The increasingly distributed nature of innovation is associated with rms linking with other rms and organisations not only on a regional and national basis but also in an international con- text. This is not a new phenomenon, especially in relation to the pharmaceutical industry (Sander- son, 1972; Liebenau, 1984), but it has continued to expand steadily. This is evident both in more traditional forms of international linkage, for example, in the growth of R&D collaboration and partnerships (see, for example, Powell, 1998; Tapon and Thong, 1999; Orsenigo et al., 2001; Hagerdoorn, 2002), as well as in newer forms (Chen, 1997), such as outsourcing of clinical trial work overseas. Increasingly, rms have extensive external research and innovation linkages form- ing complex distributed innovation networks (Coombs and Metcalfe, 2002; Chang, 2003), and leading to industries with highly open research and technical systems. This has been charac- terised by Chesbrough (2003a, b) as part of the open innovation phenomenon. What is less discussed in relation to the growth of R&D collaboration and networking is that these increasingly extensive networks incur high scanning, coordination and learning costs asso- ciated with establishing and maintaining these networks. In addition, most studies adopt a very static view of R&D linkage networks. It is not just the cost of building these networks it is their maintenance that can pose a heavy burden to rms. Some highly dynamic networks may in- volve considerable churn in network partners. The departure of a key collaborator, (a), may involve a cascade of changes with related partners of (a) departing from the network. However, although a highly dynamic pattern of network change may cause complications, equally, too little change, associated with the stasis and lock-in of research networks, can also represent a hazard. Moreover, as the number of partners increases in a network, the more likely there will be potential conict between partners within the network. Ex- clusive deals and licenses with one partner rm may exclude collaboration with another rm. Lastly, there are increased dangers of intellectual property (IP) loss with increased number of colla- borators unless rms also spend time and money monitoring and guarding against such loss. R&D networking may therefore indeed be ne- cessary (and indeed become a core competence of the rm; Madhavan et al., 1998, p. 445); however, as the cost and complexity of R&D continues to grow it may also impose an increasing resource burden to the rm seeking to master such pro- cesses. Firms need to review constantly the benets and costs of R&D collaboration and networking; it may not always be as good as it seems. 2.3. Growth and extent of external R&D activity Firms are sourcing more of their research and technical requirements externally, with their over- all level of spend increasing signicantly over the last few years (Howells, 1999; Jones, 2000; Ho- wells et al., 2003). The growth in R&D outsour- cing is also reected in the overall and relative spend by rms on outsourcing R&D. In 2003, R&D performing rms in the United States spent $10.2 billion in contract R&D to other domestic Jeremy Howells 242 R&D Management 38, 3, 2008 r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd rms. Moreover, the average annual growth rate of contracted out R&D from 1999 to 2003 was double the growth rate of in-house R&D in the United States (9.4% compared with 4.9% 3 ; Na- tional Science Foundation, 2006, pp. 433). How- ever, extramural R&D is still low for the United States at around 5.7% of total R&D in 2003 (compared with 3.7% in 1993). For other coun- tries it is much higher. Thus, data from Ireland for 2003 revealed that R&D outsourcing amounted to some h156.4 million compared with total business expenditure in research and devel- opment (BERD) of h1,075.6 million in the same year (Forfa s, 2005, pp. 7, 37), equivalent to 14.5% of the total Irish industry R&D expenditure. It should, however, be noted that R&D outsourcing here included the sourcing of R&D from other parts of the parent company (within the formal denition of the rm), but outside the business unit or enterprise in Ireland (see Section 2.6). This growth in external R&D spending can also be seen at the rm level. To take one example of a survey of R&D outsourcing in the UK pharma- ceutical industry, the number of rms outsourcing grew from 56.6% in 1998 to 71.7% in 2003, while in terms of expenditure the percentage of the R&D budget being spent externally increased from some 13.8% of the R&D budget spent externally in 1998, to 25.8% by 2003, a doubling over the ve-year period (Howells et al., 2008). Similarly, given that approximately a quarter of all pharmaceutical R&D expenditure is spent externally in Britain, this equates to a market in just one manufacturing sector in the United King- dom of some $1.5 billion (d800 million) in 2004, not a minor sum for just one industrial sector in a single country. R&D outsourcing should, there- fore, not be seen as a minor or a quaint appen- dage to the wider outsourcing debate. In terms of factors leading to such growth in external R&D, activity time, cost and lack of in- house R&D and technical expertise come out as the most highly rated factors associated with the decision to outsource. Undertaking research and technological activity in-house, or seeking exter- nal sources for it, is all about maintaining future competitiveness of the company and the future core competencies and capabilities of the rm. Other factors in the decision to source externally are the degree of task modularity and knowledge tacitness. In terms of the research or technical activity, i.e. how easy or difcult was it for the rm to decompose or separate out a particular research or technical task and to then parcel it out to another rm or organisation to undertake (see e.g. Mikkola, 2003). Thus, where a research or technical activitys boundary is very indistinct and involves closely interrelated and complex linkages with other research and technologies, outsourcing would be less likely to compare with an activity that forms a very distinct component and where its relationship with other research areas or tech- nologies is simple and clear. The degree of tacit- ness of the knowledge being transferred is also seen as a factor hindering research and technol- ogy transfer, which may in turn inuence the R&D outsourcing decision; the less explicit the know-how is, the more difcult it is to assim- ilate it (Cohen and Levinthal, 1990; Lane and Lubatkin, 1998). Clearly, the degree of tacitness is likely to affect the ability to absorb the research process back into the rm. Indeed, absorption issues, such as different research cultures between the two or more organisations involved, can be seen as a more general outcome factor inuencing the decision of whether to outsource or not. 2.4. The cost, productivity and effectiveness of R&D The late 1980s and the early 1990s marked the beginning of signicant attempts by rms to reduce their R&D expenditures. This was initially more about cost savings, less about increasing the productivity and efciency of R&D as rms in advanced industrial economies sought to reduce the overall cost base in the face of recession. Savings in R&D expenditure was of immediate benet to the rms bottom line and in the short term did not have a signicant impact on the performance of the rm. However, it was often associated with, or resulted in signicant, organi- sational change, the most crucial here being the closure of central, corporate R&D laboratories and the shift towards atter R&D structures, devolved to operational contexts lower down the organisation, namely Divisions and Strategic Business Units (SBUs). By the late 1990s, rms had moved away from simple cost savings towards improving productiv- ity by targeting R&D in terms of improving efciency. This was often associated with identi- fying and closing down unsuccessful projects ear- lier on and by fast tracking products that seem to have greatest chances of success, which has been most marked in the pharmaceutical industry (see e.g. DiMasi et al., 2003, p. 171; see also DiMasi, 2000). One of the sectors that moved early here was the pharmaceutical industry, where R&D New directions in R&D r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd R&D Management 38, 3, 2008 243 costs, especially development costs, had been spiralling but with no commensurate growth in novel drug output; indeed, productivity levels had been declining relative to outputs of new chemical entities (NCEs). Thus, fewer than 5% of all compounds that are screened enter pre-clinical development, and only 2% of these candidates enter clinical testing. Of all drugs that enter Phase I trials, approximately 80% fail in the develop- ment stage (Bolten and DeGregorio, 2002). However, these efciency and productivity im- provement initiatives remain a new phenomenon for most other sectors and rms. There remains a lot more to be done here. Interestingly, although these initiatives, in the short run, have led to a constraint on the growth of R&D expenditure, in the long run, it will improve the ability of R&D to deliver new products, processes and services and thereby encourage more R&D investment. 2.5. Globalisation of R&D and the increasing spatial division of R&D: the geographical widening and deepening of R&D The globalisation of R&D continues, but is now entering a new phase. Advanced industrial nations still perform most of the worldwide R&D. OECD member states, for example, accounted for an esti- mated 82% of the total worldwide R&D in 2000 (National Science Foundation, 2006, pp. 440). The global expansion of R&D has, therefore, been largely one of investment by multinational enter- prises (MNEs) from one advanced economy going into another advanced economy (Howells, 1990a, b). This was associated with either supply-side reasons related to asset augmentation factors (i.e. to gain access to labour, resources or know-how readily available in the parent headquarters home country), or for demand-driven asset-exploiting reasons (namely, adapting products to foreign markets and to be able to lend technical support to offshore manufacturing plants). Global expansion of R&D up until the end of 1990s could therefore be char- acterised as more of the same, in similar locations. However, we are now entering a new phase of the globalisation of R&D associated with the geographical widening and deepening of R&D in relation to the internationalisation of R&D. This has three new closely interrelated characteristics: (1) after a period of limited expansion, R&D globalisation is now entering a period with a much wider geographical span from the North to the South and from the West to the East; (2) the search for lower cost solutions via R&D offshoring (linked to productivity and ef- ciency drives in R&D; Section 2.4); and (3) the desire to get closer to consumers in the rapidly growing economies of China, India and Brazil. The complete globalisation of R&D still has a long way to go; large parts of Africa, South America and central Asia still remain off the global R&D map. However, recently, there has been a dramatic growth in R&D activity within China, India and southeast Asia. These changes are indeed being reected in the changing value placed on China and India as destinations for R&D investments in recent years. Thus, in 1994, China was only ranked 30th in 1994 ($7 million) in terms of where US rms conducted their R&D, but 11th by 2000 ($506 million; National Science Foundation, 2005, pp. 469). More specically, the most recent UNCTAD survey of the largest R&D spenders among MNEs for 2004 revealed that China was now the third largest global destination, behind the United States and the United Kingdom, and India the sixth most im- portant location (UNCTAD, 2005, p. 133). Of the 885 R&D-oriented greeneld FDI projects an- nounced in the Asian region, covered by the UNCTAD survey, three-quarters (723) were in China and India. This is in part associated with the lowering of barriers to undertaking R&D in China (Gassmann and Han, 2004). Part of this shift is linked to the phenomenon of R&D offshoring (although in reality it is difcult to distinguish this from the more general process of R&D globalisation). R&D offshoring can be asso- ciated as both an intra-rmand an inter-rmactivity (Figure 1). What has been seen as being distinct about R&D offshoring, compared with previous rounds of international R&D activity, is that: (a) it is a more cost-led process; (b) because of this, it is more closely linked to investments in less developed economies; and (c) in its inter-rm form, it is closely linked to the process of R&D outsourcing. As noted above, it is difcult to untangle R&D offshoring from the wider process of R&D glo- balisation, although at a rm level there is some evidence that MNEs are certainly themselves dening certain types of low-cost R&D invest- ment as R&D offshoring, particularly in less developed but rapidly growing economies (see Jeremy Howells 244 R&D Management 38, 3, 2008 r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd Zedtwitz et al., 2004). It is perhaps too early to be able to determine fully the phenomenon of R&D offshoring overall, and in particular in terms of cost criteria (see e.g. Jones and Teegan, 2003). However, there is evidence of close links between offshoring and outsourcing. Thus, in Ireland, of the h156.4 million of R&D activity outsourced, some h101 million, nearly two-thirds (64.6%), was spent outside Ireland. Eighty-eight percent (h138.8 million) of outsourcing was undertaken by foreign-owned companies, usually via related companies (parent, subsidiary or afliate compa- nies); this might be seen as intra-rm offshoring (Section 2.3; category 4) (Figure 1). Thirdly, there has been the rapid expansion of the economies of China, India and Brazil, coupled with the desire to get closer to consumers in relation to R&D activity (Section 3.4). Despite the many supply-led claims of R&D globalisation, the demand side is strongly correlated with the expansion of foreign R&D activities (and still associated with overseas R&D activities helping to adapt technologies and products to local market requirements). There is also evidence that shows that MNEs rarely internationalise R&D to com- pensate for technological weaknesses at home. A study by Le Bas and Sierra (2002) showed that, most often, R&D offshoring takes place in loca- tions that are strong in technology and in technol- ogies where the parent rm has an advantage. The picture of R&D globalisation is therefore changing, but is associated with a complex mix of both supply and demand factors that are also operating at a more specialised and multilateral geographical level (OECD, 2005); this is particu- larly true in relation to the offshoring of R&D (Chen, 2004). Lastly, in addition to managing (inter-rm) R&D offshoring (category 4, Figure 1) are the ongoing spatiotemporal pressures of managing R&D worldwide (Howells, 1995). Many multi- national companies are not only facing ever shorter development cycles (see e.g. Stalk and Hout, 1990) but at the same time also confronting wider geographical spans 4 of their R&D estab- lishments. For some rms, it may be a gradual process (rm A, Figure 2), but for others they are facing squeezed developments cycles at the same time as they are expanding their R&D capabilities more widely (rm B). 2.6. The growth of service R&D Service sector R&D is growing rapidly. Thus, between 1990 and 2001, service-sector R&D 5 in- creased at an average annual rate of 12% across OECD member countries, compared with ap- proximately only 3% for manufacturing sectors. Thus, by 2002, the European Union (EU) average for the share of services in BERD had increased to over 15%, although for some countries, such as the Portugal, Denmark, United States and Greece, this was much higher at over 30%. In the United States, nearly 40% of all business R&D is per- formed in the service industries, whereas in the EU this share is only 15% (Figure 3). However, since 1997, an increasing proportion of business R&D is being performed in the services sector in Europe (from 11.5% in 1997 to 15.1% in 2002). The increasing importance of services sector R&D is perceived to be due mainly to three factors (European Commission, 2005, p. 37): an improvement in the measurement of services sector R&D; a growth in R&D intensity in the services sectors; and an increase in the outsourcing of R&D by both the business and the government sectors. Obviously, within the overall services segment, there is an important sector for R&D activity the R&D services sector. This sector covers the NACE/ ISIC schema 73, namely research and development Home (Domestic) Country In-House R&D 1) Domestic, In-House R&D 3) Domestic, Outsourced R&D 4) Offshore, Outsourced R&D 2) Offshore, In-House R&D Foreign (Overseas) Country Outsourced R&D Figure 1. Typology of R&D Outsourcing and Offshoring. R&D Development Time Span (months/years) R&D Establishment Spatial Span (adjusted concentration index) 1 4 3 2 1 2 3 4 Zone of Spatio- Temporal Challenge Zone of Spatio- Temporal Comfort Firm B) Firm A) Figure 2. Stages in the SpatialTemporal Complexity of R&D. New directions in R&D r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd R&D Management 38, 3, 2008 245 services. It is a signicant sector in its own right and obviously an important generator and provider of R&D to other parts of the economy. Thus, in 2002, across the EU25, the value added by R&D services was h13.3 billion, although most of this was accounted for by the EU15 h11.9 billion. EU25 employment in R&D services was 365.5 thousand (EU15 289.7 employees). Over the next 1020 years, service R&Dis likely to become the dominant form of R&D activity in many developed economies. Increasingly, both service and manufacturing rms are undertaking or contracting in more service-like R&D as they seek to get closer to customers and gain high value added. As yet, though, very little analysis of this rapidly growing segment of the knowledge-based economy has been undertaken and both rms and nations remain ignorant around these particular changes in R&D. 3. Prospective trends in R&D 3.1. Outline This section of the paper now takes a more prospective look at R&D trends in relation to what the author sees as becoming more signicant over the next 510 years, or what is already a signicant issue but remains overlooked by man- agers, policymakers and academics. These key issues are as follows: 1. the supply and availability of R&D talent worldwide; 2. the changing nature of R&D itself its spread and intertwining with other activities; 3. the blurring between producers and consu- mers of R&D, both internal and external to the rm; 4. the control/creativity trade-off issue in R&D; and 5. the new forms of R&D organisation; and, closely linked to this; and 6. the emergence of new R&D actors. 3.2. Competing for global talent: supply and availability of high-level R&D workers Section 2.5 has already described recent changes in the globalisation of R&D, but there is one aspect of this ongoing process that deserves a separate mention. It is a basic issue, but one that nevertheless will increasingly frame and deter- mine the development of R&D over the next decade, namely the lack in the supply of talented, creative R&D workers. With the continued growth in R&D activity, there has been a growing realisation and concern about the availability of skilled R&D and technical staff due to demographic trends and patterns in university enrolment. The prospect that China and India would provide a new 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 L a t v i a S l o v a k i a P o r t u g a l C y p r u s E s t o n i a D e n m a r k U S G r e e c e C z e c h
R e p u b l i c I r e l a n d S p a i n A u s t r i a I t a l y U K H u n g a r y L i t h u a n i a N e t h e r l a n d s P o l a n d E U - 2 5
( 1 ) B e l g i u m S l o v e n i a F i n l a n d F r a n c e S w e d e n G e r m a n y J a p a n Figure 3. Share of Business Expenditure in Research and Development Performed in Services (%), 1997 and 2002. Jeremy Howells 246 R&D Management 38, 3, 2008 r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd stream of untapped R&D workers that could be used by existing large R&D players has failed to materialise as Chinese companies acknowledge they too are running out of key R&D workers. Thus, although the number of R&D workers increased from around 531,000 in 1999 to around 811,000 in 2002, the volume of R&D investments has been seen to grow even faster, with the real-term average annual R&D growth in China being 15.2% between 1991 and 2002 (UNCTAD, 2005). 6 This is also at a time when the quality of R&D supply is cited as the key locational criterion for attracting in new business investment in R&D to a locality. Thus, the lack of suitably qualied R&D personnel is also reected increasingly in R&D location decisions. It has been found that the number of university- and PhD-trained scien- tists and engineers in a locality is positively associated with foreign R&D investment in that country (Jones and Teegan, 2003). It is highly trained scientists and engineers who are therefore being sought, and who attract foreign investment in R&D. Thus, some MNEs seem to be setting up operations overseas in order to tap into knowl- edge and technology sources in centres of scien- tic excellence located worldwide. The supply of creative and science and engineering talent is going to become ever more crucial over time. 3.3. The changing nature of R&D R&D activity has been constantly evolving and changing. This, in part, has been reected in changes in the denition of R&D used by govern- ments and ofcial organisations. The most widely adopted denition of R&D is that dened by the OECD under the Frascati denition, the latest version being published in 2002 (OECD, 2002). The manual has undergone three revisions since the denition rst appeared in 1963 (OECD, 1963, 1981, 1989, 2002). Changes, associated with redenition, are therefore not a new phe- nomenon, while links with other key functions, such as production and sales and marketing, have also been acknowledged for a substantial time. However, changes associated with its expansion and blurring of R&D especially with other cor- porate activities are arguably more recent and substantial. The most signicant and earliest of these changes has been that of R&D and design, which started to emerge in the early 1990s and was associated with the shift towards electronic design tools, in particular computer-aided design (CAD), as well as the need to improve integration of R&D with other key activities. It was particu- larly visible with a number of key manufacturing sectors including the automotive sector and elec- tronics. Because of their increasingly close links, studies examining globalisation of R&D in these sectors often grouped the activities together under the heading Research, Development and Design (R, D&D 7 ; see e.g. Miller, 1994). Links with design have also been coupled with closer links downstream with computer-aided modelling, ra- pid prototyping techniques and rendering of computer graphics, in turn allowing effective simulation or emulation of key processes. Another activity that has also become more inter-linked with R&D is that of testing, mainte- nance and monitoring, which was always linked to development activities, but has become closer be- cause of advances in technology. These include rapid prototyping, Laboratory Information Man- agement Systems (LIMS) and Laboratory Infor- mation Systems (LIS), a range of electronic and design testing developments [associated with e.g. engineering valuation test (EVT) or design valua- tion test (DVT)] and remote monitoring and testing developments. Here industrial consumers are clo- sely linked with the co-development of such new pieces of equipment. A third area where R&D has also become close to other corporate functions is in the softer side of market research and strategic forecasting associated with increasingly specialised marketing and core strategy functions within the rm. This is particularly true in fast-changing, consumer-led technologies where tastes and de- mands are changing every three to six months. What does all this mean for R&D? It indicates that R&D has become even harder to dene and delimit at a time when control over these re- sources and activities, especially in terms of costs and efciency, has become ever more critical to the success of the rm. This issue is also closely associated with the growth in the co-production (and co-consumption) of R&D, which is now discussed below (Section 3.4). 3.4. Co-production and co-consumption of R&D Increasingly, activities related to research and innovation are co-produced and indeed co-con- sumed. Thus, the issue of who is producer and who is consumer of research is often becoming less clear. Even with arms length contracting out of R&D (Haour, 1992), there is a high degree of interaction and as we move through the spectrum of more informal reciprocal collaborations, there New directions in R&D r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd R&D Management 38, 3, 2008 247 is a greater likelihood of co-production in the invention and innovation process. Many users innovate (in the industrial and consumer product industries, this is between 10 and 40%), and rms that systematically involve lead-users into their R&D process can reap a considerable competitive advantage (von Hippel, 2005, p. 20). Close con- sumer interaction in the R&D and new product development (NPD) process is closely associated with project success (Salomo et al., 2003; see also Callahan and Lasry, 2004). Some of the attractions of R&D outsourcing (Howells et al., 2003) are that R&D projects are easier to manage and that many of the responsi- bilities of the actual tasks are taken away from the rm that outsources. However, it is argued here that to be truly successful in externalising R&D operations, rms will have to remain closely involved in the process as active R&D con- sumers. This may not require as much resource as undertaking the R&D in-house, but more than is commonly supposed in outsourcing other types of functions or activities. R&D remains special and indeed may be be- coming more special in relation to outsourcing activities. This, therefore, has lessons not only for the outsourcing rm but also for the supplier of R&D services. In this latter context, the R&D supplying rm may have to accept even closer collaboration with its client rms if it is to be a successful supplier of R&D services. 3.5. Control/creativity trade-off in R&D Many of the trends outlined above have suggested that rms have had to exert more control over their R&D operations than they have had to do previously. In particular, efciency (associated with the need to improve R&D productivity) and coordination (the need to manage geographi- cally dispersed units within and between the rm) has meant that control in terms of monitoring and closer supervision of R&D has increased. However, this need for more control has come at a time when many industries and rms are seeking greater creativity and entrepreneurship from their research and technical staff, as well as attracting new, highly qualied, creative per- sonnel (Section 3.5). Attempting to resolve these conicting trends is, however, a difcult task for rms. However, there are some solutions, four of which are suggested (Table 1): 1. controlled freedom; 2. time out; 3. hidden supervision; and 4. zoning of R&D. These are some initiatives that have or may emerge to resolve this control/creative trade-off but more are likely to appear in an attempt to solve this growing problem. The problem is be- coming more urgent and other solutions need to be developed to balance the conicting require- ments of successful R&D in this respect. 3.6. Emergence of new forms of R&D organisation R&D is constantly evolving in terms of the way R&D is organised. The adoption and develop- ment of matrix structure structures in R&D in the 1970s and 1980s (Katz and Allen, 1985) is one such example, while the shift away from central group R&D laboratories and R&D hierarchies in the late 1980s and early 1990s towards atter R&D structures was another. Moreover, devel- opments in information and communication tech- nologies (ICTs) have allowed for multinationals to run round-the-clock 24/7 virtual R&D opera- tions on a worldwide basis (see e.g. Howells, 1995; Chiesa and Manzini, 1997; Gassmann and von Zedtwitz, 2003) and have potentially allowed for much greater exibility in the shape and variety of organisational forms for R&D. Table 1. Managing the control/creative R&D trade-off Strategy Description 1.Controlled freedom This is by allowing a large degree of slack in time and resource allocation for staff, sometimes called harmonized management (becoming increasingly rare). 2.Time out Time out of dedicated, supervised R&D work. This may relate to 510% of total time commitments and, for a number of companies, this framework has been in place for some time (see e.g. 3M and Xerox). This is a much more regulated regime than the controlled freedom strategy. 3.Hidden supervision This is by putting in place staff within the R&D unit that seek to remove as much of the administrative burdens and responsibilities for progressing research projects to allow the targeted creative R&D staff to be creative. 4.Zoning Zoning of R&D within the rm. This is achieved by allowing some research groups, units and locations to have much looser supervisory control than others. Jeremy Howells 248 R&D Management 38, 3, 2008 r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd Indeed, rms continue to experiment with types of R&D organisation. An example is the rise of co-located research groups, involving a mix of rm and university or public research establish- ment staff; this has grown rapidly since the 1980s to enable companies to tap into and work to- gether with research groups involved in longer term basic research projects (e.g. Rolls Royce has a worldwide network of University Technology Centres). Firms have also sought to establish new organisational arrangements to help bridge the often-difcult transition between basic research and development or between R&D and produc- tion. Thus, in the former case, GlaxoSmithKline (GSK) has established Centres of Excellence for Drug Discovery (CEDD) to help bridge the gap between discovery and development activities. Table 2 seeks to take a more prospective look at what new forms of R&D unit may emerge over the next 1020 years. Some units have already emerged in their nascent form such as shared research facilities, exemplied by SEMATECH in the United States in the semiconductor re- search eld, or R&D bridging units, already highlighted in the case of GSKs CEDDs. Firms have to operate much more exible and permeable R&D structures. The R&D boundaries of the rm have become much more open and uid (Pisano, 1990). The increasing complexity of R&D organisation is, however, not just in terms of organisational forms, it is also in terms of how rms organise their human resources (HR) within the context of R&D. Companies are increasingly having to operate Extended Internal Labour Markets (EILM). These EILM are associated with a core group of R&D workers with stable jobs surrounded by a more uid periphery made up of temporary workers or those on secondment from specialist contract service rms specialising in scientic and technical services or in providing specialist contract workers (Lam, 2005). The core and periphery model is, however, a uid organi- sational structure that responds to changes in the rms labour demands. R&D staff capable of performing gatekeeping and boundary spanning functions are seen as crucial in maintaining the organisational and cultural coherence of these more uid, permeable R&D structures. 3.7. Emergence of new R&D actors Although the main growth in R&D over the past decade has been in private, business-level R&D, we should not ignore the emergence of new R&D actors within the research and innovation system. Although most studies focus on the traditional set of actors involved in R&D, such as rms, uni- versities and public or government research estab- lishments, there is an increasing number of more diverse set of more hybrid R&D organisations. These represent a much wider and more varied type of R&D facility, often involving new combi- nations of publicprivate partnerships, such as the Institute for One World Health. However, despite this growth, we remain woe- fully ignorant of what might be termed the ecology of R&D of our respective national re- search and innovation systems. We need to rectify this if we are to help manage the overall national system of innovation, as well as to help advice and support rms themselves. Table 2. New forms of R&D organisation: a prospec- tive vision New organisa- tional form Description 1. Shared laboratories Where two organizations share the same research laboratory to work on the same research topics, for example: (a) a rm and another organisation, such as university research group or public research establishment unit; or (b) two rms together (customer or supplier, for example). Large (multi-rm) shared facilities are already common, such as the creation of SEMATECH in the United States in 1987. 2. Pico R&D establishment Very small R&D unit set up in a particular part of the rm (for example, factory) to work on a specic set of research or development issues. 3. Transitory R&D unit A limited life R&D unit, similar to a pico R&D establishment, often to resolve a specic research or technical problem. 4. Automated R&D factory Virtually un-manned R&D unit undertaking highly automated research and test procedures; results are remotely monitored and supervised. 5. Consumer engaged research unit Facility to work directly with key consumers, either personal or industrial consumers (the latter case may be linked to 1 above). 6. R&D bridging unit To aid bridging the research and development divide within R&D or the R&D and production side. Becoming more established. New directions in R&D r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd R&D Management 38, 3, 2008 249 4. Conclusions What do all these current and prospective changes mean for R&D and for R&D managers? How does it help explain the paradox of continued growth in R&D despite the apparent low esteem and value placed upon R&D by rms? Undoubtedly, the worldwide growth in R&D has been spurred by the very rapid increase in research capacity parti- cularly in China and India, and also in Brazil and the smaller economies of east Asia, including Taiwan and Korea. These countries, and the com- panies within them, need R&D to meet their objectives of successful expansion and growth. More particularly, the answer lies in the fact that successful rms (and nations) still nd R&D highly valuable. For these rms (and nations), R&D still delivers in terms of generating value added in terms of new and improved products, services and pro- cesses. More particularly, at the individual project level, R&D projects meet the risk/value return that the company has specied, despite the overall scepti- cism that senior management may have about R&D actually delivering adequate results. At the national level, governments also still see R&D as an impor- tant part of the innovation infrastructure to make a national systemof innovation successful and in terms of becoming a proper Knowledge-based Economy. This is, moreover, not solely in direct terms. R&D aids the absorption into a rm or an organisation of knowledge and technology from outside. Employing R&D workers also has signicant local and regional economic and employment multiplier effects. They also have a wider cultural and social signicance. In short, R&D workers are good people to have in an area; they create signicant spillover effects in the rest of the economy. However, at both the rm and the national level, we still remain poor at adequately measuring the benets of R&D. Acknowledgements This paper arises out of funding by the UK Economic and Social Research Council (ESRC) through ESRC Centre for Research on Innova- tion and Competition (CRIC) centre-related funding (Grant Number ESRC M549285002) and Manchester Business School. The views ex- pressed are of the authors alone. References Bolten, B.M. and DeGregorio, T. (2002) Trends in development cycles. Nature Reviews: Drug Discovery, 1, 335336. Callahan, J. and Lasry, E. (2004) The importance of customer input in the development of very new products. R&D Management, 34, 107120. Chang, Y.-C. (2003) Benets of co-operation on in- novative performance: evidence from integrated cir- cuits and biotechnology rms in the UK and Taiwan. R&D Management, 33, 425438. Chen, S.-H. (1997) Decision-making in research and development collaboration. Research Policy, 26, 121135. Chen, S.-H. (2004) Taiwanese IT rms offshore R&D in China and the connection with the global innova- tion network. Research Policy, 33, 337349. Chesbrough, H. (2003a) The era of open innovation. MIT Sloan Management Review, 44, 3541. Chesbrough, H. (2003b) The logic of open innovation: managing intellectual property. California Manage- ment Review, 45, 3358. Chiesa, V. and Manzini, R. (1997) Managing virtual R&D organizations: lessons from the pharmaceutical industry. International Journal of Technology Man- agement, 13, 471485. Cohen, W.M. and Levinthal, D.A. (1990) Absorptive capacity: a new perspective on learning and innova- tion. Administrative Science Quarterly, 35, 128152. Commission of the European Communities (1989) A Framework for Community RTD Action in the 90s Discussion Document, Commission of the European Communities, Brussels. Commission of the European Communities (2002) More Research for Europe Towards 3% of GDP. COM (2002) 499 Final, Communication from the Commission, Commission of the European Commu- nities, Brussels. Commission of the European Communities (2003) Investing in Research: An Action Plan for Europe. COM (2003) 226 Final, Communication from the Commission, Commission of the European Commu- nities, Brussels. Coombs, R. and Metcalfe, J.S. (2002) Innovation in pharmaceuticals: perspectives on coordination, combination and creation of capabilities. Techno- logy Analysis and Strategic Management, 14, 261271. DiMasi, J.A. (2000) New drug innovation and phar- maceutical industry structure: trends in the output of pharmaceutical rms. Drug Information Journal, 34, 11691194. DiMasi, J.A., Hansen, R.W. and Grabowski, H.G. (2003) The price of innovation: new estimates of drug development costs. Journal of Health Econom- ics, 22, 151185. European Commission (2005) Key. Brussels: DG Re- search, Commission of the European Communities. Forfa s (2005) Business Expenditure on Research and Development (BERD), Ireland 2003/4. Dublin: Science and Technology Indicators Unit, Forfa s. Gassmann, O. and Han, Z. (2004) Motivations and barriers of foreign R&D activities in China. R&D Management, 34, 349365. Jeremy Howells 250 R&D Management 38, 3, 2008 r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd Gassmann, O. and von Zedtwitz, M. (2003) Trends and determinants of managing virtual R&D teams. R&D Management, 33, 243262. Hagerdoorn, J. (2002) Inter-rm R&D partnerships: an overview of major trends and patterns since 1960. Research Policy, 31, 477492. Haour, G. (1992) Stretching the knowledge-base of the enterprise through contract research. R&D Manage- ment, 22, 177182. Howells, J. (1990a) The internationalisation of R&D and the development of global research networks. Regional Studies, 24, 495512. Howells, J. (1990b) The location and organisation of research and development: new horizons. Research Policy, 19, 133146. Howells, J. (1995) Going global: the use of ICT net- works in research and development. Research Policy, 24, 169184. Howells, J. (1999) Research and technology outsour- cing. Technology Analysis and Strategic Manage- ment, 11, 591603. Howells, J., Gagliardi, D. and Malik, K. (2008) The growth and management of R&D outsourcing: the case of UK pharmaceuticals. R&D Management, 38, 205219. Howells, J., James, A. and Malik, K. (2003) The sourcing of technological knowledge: distributed in- novation processes and dynamic change. R&D Man- agement, 33, 395409. Katz, R. and Allen, T.J. (1985) Project perfor- mance and the locus of inuence in the R&D matrix. Academy of Management Journal, 28, 6787. Jones, G.K. and Teegan, H.J. (2003) Factors affecting foreign R&D location decisions: management and host policy implications. International Journal of Technology Management, 25, 791813. Jones, O. (2000) Innovation management as a post- modern phenomenon: the outsourcing of pharma- ceutical R&D. British Journal of Management, 11, 341356. Lam, A. (2005) Work roles and careers of R&D scientists in network organizations. Industrial Rela- tions, 44, 242275. Lane, P.J. and Lubatkin, M. (1998) Relative absorptive capacity and interorganizational learning. Strategic Management Journal, 19, 461477. Le Bas, C. and Sierra, C. (2002) Location versus home country advantages in R&D activities: some further results on multinationals locational strategies. Re- search Policy, 31, 589609. Liebenau, J. (1984) International R&D in pharmaceu- tical rms in the early twentieth century. Business History, 26, 329346. Madhavan, R., Koka, B.R. and Prescott, J.E. (1998) Networks in transition: how industry events (re)- shape interrm relationships. Strategic Management Journal, 19, 439459. Mikkola, J.H. (2003) Modularity, component outsour- cing, and inter-rm learning. R&D Management, 33, 439454. Miller, R. (1994) Global R&D networks and large-scale innovations: the case of the automobile industry. Research Policy, 23, 2746. National Science Foundation (2005) National Patterns of Research and Development Resources: 2003. Ar- lington, VA: Division of Science Resources Statistics, Directorate for Social, Behavioral, and Economic Sciences, National Science Foundation. National Science Foundation (2006) Science and En- gineering Indicators, 2006. Arlington, VA: National Science Foundation. OECD (1963) The Measurement of Scientic and Tech- nical Activities: Proposed Standard Practice for Sur- veys of Research and Experimental Development (Frascati Manual). Paris: OECD. OECD (1981) The Measurement of Scientic and Tech- nical Activities: Proposed Standard Practice for Sur- veys of Research and Experimental Development. Paris: OECD. OECD (1989) The Measurement of Scientic and Tech- nical Activities: Proposed Standard Practice for Surveys of Research and Experimental Develop- ment (Supplement to the Frascati Manual). Paris: OECD. OECD (2002) Frascati Manual: Standard Method Pro- posed for Surveys of Experimental Research and Development. Paris: OECD. OECD (2005) Internationalisation of R&D: Trends, Issues and Implications for S&T Policies. Background Report, Forum on the Internationalisation of R&D, 2930 March, 2005, Brussels. Orsenigo, L., Pammolli, F. and Riccaboni, M. (2001) Technological change and network dynamics: lessons from the pharmaceutical industry. Research Policy, 30, 485508. Pisano, G.P. (1990) The R&D boundaries of the rm: an empirical analysis. Administrative Science Quar- terly, 35, 153176. Powell, W.W. (1998) Learning from collaboration: knowledge and networks in the biotechnology and pharmaceutical industries. California Management Review, 40, 228240. Radjou, N. (2006) Does corporate R&D still matter? Research Technology Management, 49, 67. Salomo, S., Steinhoff, F. and Trommsdorff, V. (2003) Customer orientation in innovation projects and new product development success the moderating effect of product innovativeness. International Journal of Technology Management, 26, 442463. Sanderson, M. (1972) Research and the rm in British industry, 191939. Science Studies, 2, 107 151. Stalk, G. and Hout, T.M. (1990) Competing Against Time: How Time-Based Competition is Reshaping Global Markets. New York: The Free Press. New directions in R&D r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd R&D Management 38, 3, 2008 251 Tapon, F. and Thong, M. (1999) Research colla- boration by multi-national research oriented phar- maceutical rms: 19881997. R&D Management, 29, 219231. UNCTAD (2005) Transnational Corporations and the Internationalization of R&D World Investment Re- port, United Nations, New York and Geneva. von Hippel, E. (2005) Democratising Innovation. Cam- bridge, MA: The MIT Press, Massachusetts Institute of Technology. Zedtwitz, M., von Gassmann, O. and Boutelier, R. (2004) Organizing global R&D: challenges and dilemmas. Journal of Innovation Management, 10, 2149. Notes 1. All prices are in US$. Unless otherwise stated. 2. It is estimated here that there were at least 4 million people employed in R&D worldwide in 2006. In addition to the United States with 1 million, Japan had some 676,000 workers in 2001 and China some 811,000 in 2002 (although OECD estimate this to be only 443,000). 3. See Commission of the European Communities (2002; 2003) 4. After adjusting for ination. 5. As measured by mean spatial concentration indexes. 6. Covering ISIC categories 5099. 7. This, however, has been a concern as far back as the late 1980s when the Commission of the European Communities (1989) estimated that the United States would face a shortfall of 500,000 scientists and engineers by 2010. 8. This is being extended further with, for example, the European Commission now using the term Research, Development and Innovation (R, D&I). Jeremy Howells is Professor and Executive Direc- tor of the Manchester Institute of Innovation Research (MIoIR) and Head of the Innovation Management and Policy (IMP) Division at Man- chester Business School, University of Manche- ster. He received his PhD from the University of Cambridge. He has published in the International Journal of Technology Management, Pro- metheus, R&D Management, Research Policy, Science and Public Policy, Service Industries Journal and Technology Analysis and Strategic Management. His current research interests are on R&D outsourcing and offshoring, industry- academic links, service innovation and technol- ogy transfer. His research has been funded by UK Economic and Social Research Council (ESRC), the Engineering and Physical Sciences Research Council (EPSRC), the European Commission, OECD, the European Science Foundation, UNIDO and UNCTAD. Jeremy Howells 252 R&D Management 38, 3, 2008 r 2008 The Author Journal compilation r 2008 Blackwell Publishing Ltd