research policy - Science Direct

research policy ELSEVIER

Research Policy 26 (1997) 121-135

Decision-making in research and development collaboration Shin-Horng Chen Chung-Hua Institution for Economic Research, 75 Chang Hsin Street, Taipei, Taiwan

Accepted 29 November 1996

Abstract Research and development (R&D) collaboration has come to the forefront of both academic interest and public policies. This paper aims to contribute to the current understanding of R&D collaboration by exploring issues regarding its decision-making aspect and drawing implications for the public policies. Specific inquiries are made into issues related to major types of research activities conducted by joint R&D partners and to where decisions on R&D collaboration are taken within multi-site corporations. The research draws on an original questionnaire survey addressed to establishments in three industries in the UK. The empirical results suggest that joint R&D activities tend to be competition related. On this basis, the paper critically examines the debate on government support for pre-competitive vs. near-market collaboration. The evidence also shows that corporate control for R&D collaboration can be decentralized. As a result, a more flexible view of the strategic nature of R&D collaboration needs to be acknowledged. Further implications can be drawn with reference to the concept of 'additionality' which has been used in rationalizing public support for R&D. © 1997 Elsevier Science B.V.

1. Introduction Since the 1980s, an increasingly significant development on the technical landscape has been the growth of research and development (hereafter R & D) collaboration. There is evidence from a number of studies (e.g. Chesnais, 1988; Hagedoorn, 1990) to suggest the existence of such a trend. In addition, the public sector has orchestrated numerous publicly funded technology programmes at various spatial levels to promote inter-organizational technological collaboration and economic development (Sharp and Shearman, 1987; Rothwell and Dodgson, 1992). As a result, R & D collaboration has come to the forefront of both academic interest and public policies. This paper aims to contribute to the current under-

standing of R & D collaboration by exploring issues regarding its decision-making aspect and drawing implications for public policies. Specifically, inquiries are made into issues such as what the major types of research activities conducted by joint R & D partners are, and where decisions on R & D collaboration are taken within multi-site corporations. Most public collaborative research programmes, except for a few cases such as the Eureka Programme, are officially claimed to promote merely 'pre-competitive research', as compared with nearmarket research. An underlying reason, it is argued, is that finns are not interested in near-market collaboration (see Dodgson, 1993a, pp. 171-172). Along this line, Ouchi and Bolton (1988) have developed a logic of joint R & D to argue that R & D collaboration

0048-7333/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved. PII S0048-7333(96)00919-5

122

S.-H. Chen / Research Policy 26 (1997) 121-135

is initiated by firms to undertake research characterized by certain types of intellectual property rights. How well do such arguments stand up to empirical scrutiny? This question is addressed in the paper through examining the major types of research activities conducted by linked partners in their collaborative process. The issue of the locus of control for R & D collaboration within corporations has been a largely neglected area of research. This may be due to a tendency to see R & D collaboration as a part of overall corporate strategy (Porter and Fuller, 1986). In light of the long-standing debate on centralization vs. decentralization in R & D activities, the paper, however, suggests a need to keep a more open mind about the nature of the decision-making process in R & D collaboration and where the decisions are taken within multi-site corporations. An exploration of the locus of corporate control for R & D collaboration is therefore needed, the results of which may facilitate our understanding of the wider nature of such ventures. As to its empirical research, the paper draws on an original questionnaire survey addressed to establishments within three industries in the UK. The empirical results suggest that joint R & D activities tend to be competition related. On this basis, the paper examines the debate on government support for pre-competitive vs. near-market collaboration. The evidence also reveals that corporate control for R & D collaboration can be decentralized. As a result, a more flexible view of the strategic nature of R & D collaboration needs to be acknowledged. Further implications can be drawn with reference to the concept of 'additionality', which has been used in rationalizing public support for R&D. The remainder of the paper is arranged as follows. Section 2 brings together and reviews the relevant but contradictory views on the two issues concerned. The research methods used in the paper are described in Section 3. Following this, empirical data on the major types of joint R & D activities are presented and discussed in Section 4. Section 5 presents evidence to show that corporate control for R & D collaboration can be decentralized. The underlying reasons are explored in Section 6. Finally, the paper draws policy implications and conclusions.

2. Two aspects of decision-making in R & D collaboration Over the last decade, R & D collaboration has become a subject of considerable academic interest. From a theoretical perspective, collaboration itself represents a departure from the well-elaborated alternative modes of organizations: market and hierarchy (i.e. the firm) in neo-classic economics. In the context of R & D , this option goes against ~he conventional view that firms are better off if they organize R & D internally, to avoid contractual difficulties in contracting out R & D (Teece, 1988, pp. 259-262) and to safeguard competitive advantages. Set against this reasoning, R & D collaboration often involves sharing knowledge with other firms and contracting with outsiders. It is, therefore, not surprising that the literature concerned has paid much attention to the reasons why firms enter into R & D partnerships (for a review, see Dodgson (1993b)). However, an under-explored theme in conjunction with R & D collaboration seems to be its decision-making aspect, which covers a wide range of issues on how firms conduct R & D partnerships. Such an issue as the selection of joint R & D partners has recently been investigated by Hakanson (1993) and Chen (1994). This paper is concerned with two questions. First, what types of research activities are actively conducted in the joint R & D process? Second, where is the locus of corporate control for such ventures? Admittedly, there is no shortage of opinions on these two issues, but they are so patchy that the relevant empirical literature remains sparse. This section serves to bring together and critically review the relevant but contradictory views. Ouchi and Bolton (1988) provide one of the several discussions about the major types of R & D activities conducted under collaborative agreements. They contend that the logic of joint R & D can be explained in terms of the nature of intellectual property. For Ouchi and Bolton (1988), R & D collaboration and other research mechanisms (the firm, university or government laboratory) map on to various types of intellectual properties in a discriminate way, reflecting problems of appropriability. Intellectual properties (private, public and leaky) differ in their degree of appropriability. The fit between institu-


tionai form and intellectual property occurs because the degree of appropriability matches the incentives for invention of each institutional form. Ouchi and Bolton argue that the private firm (i.e. internal R & D ) is the most suitable form of institution to generate private property which can be safeguarded by a private party. Owing to its nature as public goods, public property should and will be generated in public institutions such as universities or government laboratories. Industrial R & D collaboration provides the institutional form appropriate to produce leaky property, which can only be privately appropriated for a short period. Such arguments have been used to justify the need for the public sector to support only 'pre-competitive' research collaboration. Normally, pre-competitive research is related to basic research (Dodgson, 1993a, pp. 171-172; Garnsey and Moore, 1993), though leaky property, for Ouchi and Bolton, appears to be applied research. However, private firms are assumed not to be interested in near-market collaboration because collaboration as such involves competitive conflicts or because it is in firms' best interest to undertake near-market research internally (see Dodgson, 1993a). Contrary to this view, Fransman (1990) argues: " I f research is to be judged to be worth undertaking for a for-profit company, it must be expected, even with uncertainty, to eventually yield competitive advantage. From this point of view, the notion of pre-competitive research in for-profit companies is a contradiction in terms" (p. 282). The argument surrounding pre-competitive research also fails to take into account the fact that the process of innovation may be an iterative one; as a result, even private collaboration can involve activities beyond basic or applied research (Jorde and Teece, 1990; Dodgson, 1993a, p. 172; Garnsey and Moore. 1993). In addition, firms may opt for collaboration for the very basic reason of seeking external help to complete vital R & D projects. It is, therefore, implausible that collaborating firms will simply limit themselves to pre-competitive research. It is fashionable to associate R & D collaboration with the term 'strategic alliance'. What this seems to imply is that a decision to collaborate in R & D is made exclusively at the corporate headquarters level.

123

Porter and Fuller are concerned with global competition and contend that "coalition can only be understood in the context of a firm's overall global strate g y " (Porter and Fuller, 1986, p. 316). At the heart of international competition are the decisions associated with configuration and co-ordination of valuechain activities (Porter, 1986). Coalitions are seen by Porter as the way by which the partners work together and configure value-chain activities on a global scale to gain competitive advantages. Similarly, Delapalme (1987, p. 17) argues that R & D collaboration must be part of strategic alliances. In this regard, R & D collaboration is reasoned to derive from a corporate perspective, such as the need for technological diversification and to block competition (see Dodgson, 1993a, pp. 30-36). However, in the context of multi-site corporations, a central and continuing debate surrounding the management of R & D is that on centralization vs. decentralization (Steele, 1975). This debate is widely known, but few commentators appear to have thought through its implications for R & D collaboration, which may form part of corporate R & D activities. Whether a multi-site firm should centralize or decentralize its R & D decisions involves a number of issues. Charles (1990, pp. 103-111) has summarized the factors concerned as follows. First, the mismatch between information and decision-making locations may lead to decentralizing R & D decisions. To formulate a coherent R & D decision within any technological strategy often requires information that originates at the bottom of the organizational pyramid. Information flows in a hierarchically complex firm are, however, not frictionless. The central part of the firm may not always hold sufficient and appropriate information underpinning decisions for R & D undertaken elsewhere in the organization. The firm, therefore, may have to decentralize its decision-making apparatus to ensure that R & D decisions are made by those who have adequate information and incentives. Furthermore, this tendency to decentralize appears to be strengthened by internationalization of R& D, especially when the offshore R & D units are the legacy of acquisition and mergers (Hakanson, 1990, p. 265). Charles' second argument is that the nature of corporate product lines, classified by a dichotomy of demand- and supply-led, may also influence the arrangement of R & D structures (also Steele, 1975,

124

S.-H. Chen // Research Policy 26 (1997) 121-135

p. 127). Supply-led R&D is characterized by internal orientation and hence may be amenable to the principles of centralization. By contrast, demand-led R&D requires more impetus from outside the R&D unit itself and hence may be better governed from a decentralized position. The third influential factor as noted by Charles is the degree of interdependence between divisions required for organizational achievement, where internal interdependence has been classified by Thompson (1967) (quoted by Charles (1990), p. 110) into pooled, sequential and reciprocal interdependence. The degree of interdependence increases with the progression from pooled through sequential to reciprocal inter-dependence, and with organizational complexity. Therefore, organizational control in R&D is likely to shift away from decentralization towards centralization as a corporation becomes more internally interdependent. In contrast to a centralizing tendency and in parallel to the theoretical reasoning, there appears to be a tendency towards market mechanisms being increasingly applied within the boundary of a firm (e.g. profit centres) to increase the efficiency and effectiveness of the hierarchy (Li, 1993, p. 51). This transformation is also found to take place in the area

of R&D management (Whittington, 1990). It has been customary to assume that divisions and/or product plants of a corporation will contract with their central R&D laboratory, if any, for R&D projects beyond their own technological capacities. According to Whittington, however, some large corporations have recently changed this convention by introducing a customer-contractor principle to the management of the linkage between the central R & D facility and its internal customers. This development involves "the conversion of technical centers into profit centers [and] trading with operating companies within the divisionalized corporate structure at market prices" (Whittington, 1990, p. 192). A consequence of such corporate transformation is decentralizing business responsibility and escalation of the autonomy of low-level tiers. The literature on R&D partnerships involved in the centralization vs. decentralization debate has tended to focus specifically on globalization issues. Despite sharing a common view of globalization, Porter and Fuller (1986) and Sigurdson (1990) (following Westney (1990)) differ in their opinions regarding the locus of control for R&D collaboration in such situations. For Porter and Fuller, a decision to forge R&D collaboration is the privilege of multi-

Strategicalliances Joint research Headquarters ( HQ ) ( HQ country)

I

I Universities

I

other company

CTU

G

Production Major Company ( Division) Domesticcountry Subsidiary ( Marketcountry)

I Production

IuTT~

Abroad I L__

[

I Research Institutes

G Production

ProductionlTU ~ ProductionlTU

I I I I I Oniversitios I I ResearohIostitos I Fig. 1. The internal and external linkages in the overseas R&D units of the multinational. Source: Sigurdson (1990, Fig. 9.3, p. 184). Reproduced with permission of Pinter Publishers.

S.-H. Chen /Research Policy 26 (1997) 121-135

national headquarters, to achieve strategic co-ordination globally. By contrast, Sigurdson (1990) and Westney (1990) emphasize the need for diversifying external R & D linkages across hierarchies within multinational corporations (MNCs) to achieve technological and economic aims. In particular, Westney (1990, p. 283) identifies four types of foreign R & D units: 1. Technology Transfer Units (TrUs): to facilitate the transfer of the parent's technology to the subsidiary, and to provide local technical services. 2. Indigenous Technology Units (ITUs): to develop new products for the local market, drawing on local technology. 3. Global Technology Units (GTUs): to develop new products and processes for world markets. 4. Corporate Technology Units (CTUs): to generate basic technology for use by the corporate parent. She then argues that "each type of unit has distinctive linkages with the local subsidiary, the parent organization, and with local sources of technology" (Westney, 1990) (Fig. 1). In terms of the ties with the local scientific and technical community, the linkages are virtually non-existent for a TTU, and become stronger as an offshore R & D unit moves from ITU through GTU to CTU. For Westney and Sigurdson, internationalization of R & D is not merely a process of relocating the R & D function on a global scale. Instead, it is a means by which MNCs tap relevant technical and market information which may emerge outside the domestic boundary (De Meyer, 1993). The above debate about corporate control of R & D collaboration emphasizes the multinational context. It, however, should be asked to what extent joint R & D involves a centralized decision in a national or subnational setting. This question is important in light of the policy trend of the promotion of co-operative R & D (Norris, 1985; Rothwell and Dodgson, 1992), science parks and technopolises (Charles, 1989) as local strategies designed to establish regional technological networks. In the UK context, the issue is particularly important, as designated science parks are concentrated in the northern or peripheral regions (Massey et al., 1992, p. 15), where we see fewer company headquarters (Evans, 1973; Goddard and Smith, 1978).

125

3. Research methods

The empirical research of the paper draws on part of an original questionnaire survey addressed to establishments in three selected sectors in the UK. The surveyed sectors, in order of industrial R&D intensity, are the instrument and control, electrical equipment and machine tool industries. (The average industrial R & D intensity over the period 1985-1989 is 16% for the instrument and control industry, 2.03% for the electrical equipment industry, and 1.43% for the machine tool industry.) By establishment, we mean industrial plant, which is either a single-site firm or a function site of an enterprise or a firm. The focus on the establishment level was driven originally by an intention to examine the characteristics, spatial patterns in particular, of R & D collaborators (Chen, 1994). Also, it is likely that R & D collaboration can take place at different levels of a multi-site enterprise. The head office therefore may not have sufficiently detailed information on the topics that are the focus in this paper. For example, if the collaborative projects are executed at sites other than the headquarters of a multi-site company, the latter may not be able to provide accurate information on the types of joint R & D activities being conducted. In addition, as shown below, the finding that corporate control of joint R & D can be decentralized further strengthens the position of focusing on the establishment as the unit of inquiry. The survey was mailed in 1992 to 1000 establishments, which were selected without prior knowledge of whether or not they were engaged in collaborative R & D ventures. In total, 229 usable questionnaires were received. Among them, 84 were returned by collaborators. The sample proved to be biased towards larger establishments. Taking into account the fact that formal R & D tends to take place at larger sites (Thwaites et al., 1981, p. 19), the bias was thought not to seriously undermine the validity of the survey and the subsequent analyses, because the bulk of R & D effort in the surveyed sectors may be reasonably represented. For the purpose of this paper, the managing director or R & D manager of each collaborator was asked to identify the types of R & D activities conducted under the establishment's major and latest collaborative R&D venture. In addition, collaborators which

126


were corporate subsidiaries were asked to identify the lowest level of hierarchy within their corporations where the managers can conclude joint R&D agreements independently. By R&D, we refer to the definition of the 'Frascati Manual' (OECD, 1981), which classifies R & D activities into basic, applied research and experimental development (hereafter development). It is known that there are criticisms levelled against the Frascati definitions (House of Lords, 1990). Keeping in mind their drawbacks, the definitions are still used in this study, because they are widely known and accepted. Obviously, we had to rely on respondents' judgement to identify the types of R & D activities under their partnerships, but they were provided with a reference note of the R & D definitions. Reasonable confidence can be placed in the accuracy of the data on joint R&D activities, given the fact that the majority of the responses were completed by the R & D managers or technical director of surveyed firms, who is arguably familiar with the Frascati definitions.

4. R&D activities of collaborative R&D projects This section examines the major types of research activities conducted by joint R & D partners. By types of research activities, we mean basic research, applied research, development (OECD, 198 l) and commercial production. Table 1 summarizes the results. It should be noted at the outset that there are signs that the multi-stage joint R & D process is not necessarily a linear one of basic research leading through applied research and development to commercial production. For instance, a respondent indicated that their collaborative R & D project involved only basic research and commercial production. There was also a joint R & D project dealing only with applied research and commercial production. The linear model of R & D (or innovation) assumes a unilateral and sequential process going from basic research, to applied research, development, and then to commercial production. It ignores any feedback or overlap between and among the stages (Jorde and Teece, 1990). In support of a non-linear model, Surrey and Chesshire (1972, p. 21) have drawn attention to the fact that new achievements in the technical parameters in electrical equipment sometimes require gaining a better understanding of basic

science. Taking this together with the findings discussed above, a possible deduction is that a joint R & D project may be driven initially by a need to solve problems arising from a downstream stage which then requires undertaking upstream research. For the collaborative project to bear fruit, it may be imperative for the linked partners to jointly undertake both the upstream and downstream research. By implication, it is unlikely that firms involved in R& D partnerships will confine themselves merely to upstream pre-competitive research, even if some public sector programmes would appear to support only such activities. Indeed, it is evident from our research that the major focus of collaborative R&D projects is development, followed by applied research and commercial production. In contrast, R&D collaboration is involved least in basic research. Among the 80 projects cited by respondents, 59 (73.75%) projects involve development, 39 (48.75%) applied research, and 36 (45%) commercial production. Only eight agreements (10%) are related to basic research. Differentiation by sectors of industry also shows that joint R&D projects are least oriented to basic research. For example, in the instrument and control industry, where we see the highest proportion of partnerships involving basic research, the percentage is only around 12%. By implication, joint R&D tends to be competition related. The development orientation of R&D collaboration becomes more apparent as one examines private projects more closely. A partnership is defined as a private project if it is neither initiated nor sponsored by the public sector. Table 1 shows that private partnerships are oriented mainly to development and commercial production, whereas the public programmes involve relatively more applied and basic research. Moreover, it is evident that no private partnerships in the machine tool sector are related to basic and applied research, which is in line with the strong development orientation of R & D in this sector. (Among the three sectors, R&D in the machine tool industry is the most development oriented, with 95.18% of the sector's R & D investment in 1989 devoted to development. The development orientation of R & D in the electrical equipment and instrument and control industries in the same year was 90.77% and 87.70% respectively.)

S.-H. Chen / Research Polio" 26 (1997) 121-135

127

Table I Research activities of collaborative R & D agreement a (values are numbers of projects, with percentage of project number given in parentheses) Sector

No. of projects

Basic research

Applied research

Development

Commercial production

Machine tool

Public: 8

l (12.50)

5 (62.50)

5 (62.50) 9

I (12.50) 7

(90.O0)

(70.00)

1 (5.56) 1 (11.11) 1 (8.33) 2 (9.52) 4 (28.57) 1 (3.70) 5 (12.20) 6 (19.35) 2 (4.08) 8 (10.O0)

5 (27.78) 7 (77.78) 3 (25.00) 10 (47.62) 8 (57.14) L6 (59.26) 24 (58.54) 20 (64.52) 19 (38.78) 39 (48.75)

14 (77.78) 6 (66.67) 7 (58.33) 13 (61.90) 9 (64.29) 23 (85.19) 32 (78.05) 20 (64.52) 39 (79.59) 59 (73.75)

8 (44.44) 2 (22.22) 10 (83.33) 12 (57.14) 3 (21.43) 13 (48.15) 16 (39.02) 6 (19.35) 30 (61.22) 36 (45.00)

Private: 10 Subtotal: 18 Electrical equipment

Public: 9 Private: 12 Subtotal: 21

Instrument and control

Public: 14 Private: 27 Subtotal: 41

Total sample

Public: 31 Private: 49 Total: 80

" Multiple answers were possible.

It is perhaps surprising to find that a considerable portion of the public joint R & D programmes identified in Table 1 involve development and even commercial production. The reasons for this could be twofold. First, some firms participating in the public programmes may have privately extended their joint efforts beyond basic and applied research to the downstream near-market activities. Second, a few recently launched public programmes, such as the Eureka Programme, have been directed at promoting near-market collaboration. It is also possible that the identified public programmes may include initiatives of local governments which appear less likely to focus on pre-competitive research. These findings contradict the 'logic of joint R & D' of Ouchi and Bolton (1988). In their scenario, .joint R & D is tenable only if the project in question involves 'leaky intellectual property', which, by their definition, is associated mainly with applied research. For-profit firms, it is argued, will not collaborate in R & D for the purpose of undertaking devel-

opment and commercial production because 'private property', which characterizes the two technical activities, can be effectively appropriated by means of internal R&D. The opposite, however, is the case, according to the evidence from this study. Private partnerships do involve and favour development and commercial production. The arguments of Ouchi and Bolton are therefore rejected or seen at least as only applicable in selected circumstances. However, a pattern seems to emerge which suggests that as an industry becomes more research intensive, its constituent firms are likely to collaborate more in applied and basic research. Table 1 shows that the percentage of projects involving basic research is 5.56% for the machine tool sector, but rises along with industrial R & D intensity, to 12.20% for the instrument and control industry. This is also the case for applied research. The contrast is more apparent in terms of the portion of private partnerships involving applied research. None of the private partnerships in the machine tool sector are formed to

128

S. -H. Chen / Research Policy 26 (1997) 121-135

undertake applied research. In contrast, joint R & D involving applied research accounts for 25% and 59.26% of the private partnerships in the electrical equipment and instrument and control industries. This pattern may be attributed to a relatively strong interaction between scientifically related research and the advance of high-technology sectors. Technological advance in the high-technology sector is arguably so dynamic that the constituent firms need constant injection of the fruits of scientific research to remain innovative. Specifically, R & D activities in the machine tool sector have been concentrated in development rather than research. Science-related knowledge is, however, both inputs and outputs of innovations in the instrument and control industry, it is, therefore, not surprising to see that R & D partnerships in high-technology sectors, such as the instrument and control industry, focus relatively more on basic and applied research. On balance, it could be argued with some validity that private partnerships, in high-technology sectors particularly, may involve applied and even basic research, but more importantly, near-market collaboration can and does exist.

5. Locus of corporate control for R & D collaboration Having explored the types of R & D undertaken by collaborators in the sectors surveyed, the paper now investigates the control of collaborative ventures by examining the lowest level of hierarchy within corporations where the managers can on their own responsibility conclude joint R & D agreements. Corporate hierarchies are divided into individual establishments, product divisional headquarters, regional headquarters and group headquarters (Charles, 1990, p. 107). Not every multi-site corporation has all these levels, but individual establishments and group headquarters are generally applicable. Table 2 summarizes the locus of control for R & D collaboration within the multi-plant companies surveyed. The responses from single-site firms were excluded from this analysis because the centralization vs. decentralization debate is inapplicable. Among the sub-sample of 68 group establishments, six pairs of establishments have the same parents.

This raised the problem of sample independence. As a result, one may overestimate the number of incidences for some categories of hierarchies which control R & D collaboration. To avoid this problem, the responses from establishments with the same parents were not double-counted. The total number of observations was therefore reduced from 68 to 62 (each representing a different firm). Although there is one case where a large UK corporation operates in both the electrical equipment and instrument and control industries, its effect on bias is negligible. There is an observable pattern of decentralizing control of R & D collaboration. Among the 62 observations, 33 parents have authorized their subsidiaries to conclude joint R & D agreements if and when they so wish. Decentralizing control is particularly noticeable in the instrument and control and electrical equipment industries. Twenty out of 31 subsidiaries in the instrument and control sector claimed that their parent companies have delegated the power concerned to individual establishments. In the electrical equipment industry, this is the case in nine out of 17 establishments. In contrast, although four subsidiaries in the machine tool sector are able to conclude joint R & D agreements on their own, the dominant locus of control in this sector is group headquarters (eight cases). However, for the 33 corporations having delegated the power in question to individual establishments within the group, it might be an exaggeration to say that every establishment within them enjoys the same degree of privilege. This qualification becomes apparent when one cross-checks the responses from those affiliates with the same parents. Four pairs of them gave identical answers to the question and all claimed that individual establishments were the locus of control for joint R & D in their corporations. However, two pairs of sister establishments responded inconsistently. Of note is the case where one respondent gave the reply that individual establishments were the locus of control, but its sister establishment considered that the locus of control was the regional headquarters. This inconsistency may be due to a difference in corporate status between the two sister establishments. One establishment might be assigned more important missions and hence enjoy greater autonomy, whereas the other could be a low-profile subsidiary and hence be sub-


129

Table 2 Locus of control for R & D collaboration within corporations (values are numbers, with number of foreign-owned subsidiaries given in parentheses) Locus of control for R & D collaboration

Individual establishments Product division headquarters

Sector Machine tool 01 = 14) (5)

Instrument and control 01 = 31) ( l 1)

4

9

20

33

(l)

(4)

(6)

(I I)

2

2

7

I1

(5)

(5) 3 (2)

Regional headquarters Group Headquarters

Total ( n = 62)

Electrical equipment (n = 17) (6)

3 (2) 8

,~ b

(4)

4

(22)

14

(4)

~ One of the respondents did not answer this question. b These two respondents particularly specified company board.

ject to supervision from higher levels in the hierarchy (a 'performance' plant versus a 'branch' plant, in Amin's terms (Amin, 1993, pp. 2-3)). Thus, individual establishments in a decentralizing corporation may not necessarily share the same degree of autonomy regarding joint R&D. None the less, it remains safe to say that a relatively high portion of group establishments claim that they are able to conclude joint R & D agreements independently. Of note is that this pattern of decentralizing control appears to apply equally to home- and foreignowned affiliates. In the sample, 22 of the group establishments are foreign owned. Among them, 11 claimed that individual establishments within their corporations were empowered to conclude joint R & D agreements, whereas four affiliates were subject to control from their overseas parents. By sector, the majority of the foreign-owned affiliates in the electrical equipment (four out of six) and instrument and control (six out of 11) industries are sufficiently autonomous to forge R & D collaboration independently. None of the subsidiaries in these two sectors needed a mandate from their overseas parents to conclude agreements. In contrast, four out of the five toreign-owned affiliates in the machine tool sector have to gain permission from group headquarters to conclude a joint R & D agreement. It is instructive to look at the nationalities of these foreign-owned subsidiaries. For subsidiaries whose overseas parents have decentralized control, two are

Swedish, one Dutch, one German, two Swiss, four are US and one is Norwegian. Among these national affiliations, companies from Switzerland, Germany and the Netherlands, and to a lesser degree the USA and Sweden, are generally marked by higher levels of global sourcing than companies from other nations (Hakanson, 1990; Casson and Singh, 1993, p. 96). Among the rest in the sample, four are owned individually by parents in less internationalized countries--Austria, Canada, Ireland and Taiwan (Casson and Singh, 1993). Whereas the remainder belong to countries that have greater relative internationalization-the USA (three cases), France (two cases), and the Netherlands (two cases)--they need to report only to their divisional or regional headquarters to conclude joint R & D agreements. It is noted that this sub-sample of foreign-owned subsidiaries remains too small to support a strong conclusion. However, the overall findings seem to suggest that affiliates owned by parents in those countries with a higher degree of internationalization tend to enjoy more autonomy in forging R & D collaboration than do those establishments associated with countries with a more limited international outlook. To summarize, it was found that the majority of corporations involved in the electrical equipment and instrument and control industries had empowered their individual establishments to conclude joint R& D agreements, whereas there were a few such cases in the machine tool sector--the dominant locus of

130


the control in this sector was at the group headquarters level. Similar patterns were also observable in so far as foreign-owned subsidiaries were concerned. It was also found that foreign parents in internationalized countries might be more likely to delegate R & D autonomy to their UK subsidiaries than were those in countries with a lower international profile.

6. Reasons for decentralized control of R & D collaboration

To explain the findings stated above, it is important to take into account the industrial features of R&D, rather than to consider R & D collaboration merely in the context of corporate strategy, a practice that has been adopted by some researchers studying this subject. It is possible that trends towards internationalization may result in decentralizing control of R & D collaboration (Sigurdson, 1990; Westney, 1990). It is generally agreed that the machine tool industry is not a highly internationalized sector (Mayes, 1991, p. 388; Young, 1990; Young et al., 1992, p. 24). Here, internationalization does not mean exports but production organized at a multi-national scale. Because of this, foreign-owned machine tool subsidiaries in the UK are unlikely to have a high degree of autonomy; therefore, their decisions to collaborate in R & D are controlled mainly by overseas parents. In contrast, the electrical equipment sector, throughout its history, has been an internationally oriented industry (Newfarmer, 1980; Thomas and McGowan, 1991). A small group of leading producers in the sector, such as GE, Westinghouse, Siemens and GEC Alsthom, are large and highly diversified across national and product markets. Even within a nation, these producers may maintain a number of establishments across regions, with some of them being equipped with R & D units. This sector's high internationalization, together with its technical and capital intensity, has shaped distinct business practices which may affect the practice of corporate control of R & D collaboration. First, as the leading firms are highly diversified, into unrelated as well as related sectors, decentralized control is likely to prevail in such conglomerates (Charles, 1990, p. 110). Second, owing to capi-

tal and technical intensity, electrical equipment suppliers sometimes have to form consortia to present packaged bids for larger systems. To grasp business opportunities of which development work is probably characterized by modifying here and adjusting there to accommodate final users' needs, group establishments in this sector may be given a relatively free hand to conclude collaborative agreements. Although the extent to which the instrument and control sector is internationalized is unclear, there is a sign that foreign-owned affiliates in this sector play a significant role in the UK market (Rendeiro, 1985, p. 186). Available evidence suggests that some of the foreign-owned establishments in the sector are fairly autonomous in terms of undertaking technological innovations. Early research undertaken in SPRU (quoted by Rendeiro and Shepherd (1981), p. 26) revealed that UK innovations in instruments introduced by both domestic and foreign firms relied mainly on in-house knowledge, aided by a certain amount of outside inputs but without contributions coming from parent companies. This may suggest that foreign (as well as domestic) group plants in the UK instrument industry are likely to act more or less independently. Three factors could be the underlying reasons for this. First, some instrument establishments are part of large diversified corporations, which may be more receptive than smaller, simpler companies to decentralizing control. Furthermore, the nature of the products in this industry are also extremely diverse and do not lend themselves to centralized development across a large multi-product firm. Second, as the instrument industry tends to be an early user of technical advances developed elsewhere (Rendeiro, 1985, p. 148), foreign-owned subsidiaries in the UK may have as part of their routines or commission to act as a 'listening post' for technological opportunities in the host country. To fulfil this task, they may be allowed to act independently to develop local linkages in the UK (Sigurdson, 1990; Westney, 1990). Third, the sector is also marked by the end users playing an important role in stimulating and facilitating innovations (Von Hippel, 1976; Rendeiro, 1985, pp. 165-166). As a result, decentralized control may be needed to effectively bridge the connection with these users. Some factors which were discussed above as sup-

S. -H. Chen / Research Policy 26 (1997) 121-135

port for the arguments surrounding the effects of internationalization on control loci are also valid in explaining the same pattern of decentralized control in the context of UK-based firms. For example, UK subsidiaries in the electrical equipment industry are possibly able to act independently to form consortia with other suppliers to present packaged bids. To take advantage of feedback from users, who play an all-important role in the instrument industry, the sector's UK-owned establishments may also enjoy some degree of autonomy which allows them to conclude joint R & D agreements on their own. The study by Charles (1990) on the strategies adopted by major UK electronic corporations such as Marconi and Racal has highlighted their increased orientation towards niche markets, in particular those associated with the defence industry. As a result, "[their] individual SBUs (strategic business units) have experienced a high degree of financial accountability to the extent of developing their own long term technological and investment strategies within a financial planning framework set by the overall company groups. Intra-corporate integration therefore has been low, the main incentive for those links that exist being project related rather than cost-induced ... Hence, due to the collaborative nature of many defense contracts, particular subsidiaries may be more closely linked into other firms than those units within lheir own group" (p. 374). His observation seems to fit reasonably well the instrument and control industry surveyed in this study, because it is part of the electronics sector and highly defence related, and some collaborators in this sector in our surveys were involved in the defence market. In addition, three more general factors can account for the pattern and degree of decentralized control for R & D collaboration. The first factor relates to co-ordination problems associated with R & D within multi-site corporations (Casson and Singh, 1993). Where a corporation has product R & D centres without a central R & D laboratory, individual R & D establishments hold and specialize in distinctive technological competencies. Specialized technical problems unsolved by one R & D unit may not fall within the technical competencies of other units, and may therefore call for collaboration with out-

131

siders. Related to this, it is arguable in the context of the internationalization of R & D that the main supplier of new technology in the firm may not necessarily be the central R & D laboratories, as "some of them are just 'first amongst equals' in a network of laboratories" (Casson and Singh, 1993, p. 103). Nor can the hierarchical notion of superior home vs. 'inferior overseas' research capacity hold in every case and be beyond doubt (Howells, 1990, p. 508; also Hakanson, 1990). A second factor underlying decentralized control of R & D partnerships is associated with a recent change in R & D management within corporations. As is noted above, Whittington (1990) has observed a tendency for some corporations to change their internal modus operandi of R & D by employing a contractor-customer principle to regulate the relationships between their central R & D laboratory and other establishments. This development involves the conversion of technical centres into profit centres and an increase in the autonomy of low-level tiers in controlling their own R & D requirements. In these circumstances, individual establishments are arguably allowed to choose for themselves between contracting with their central R & D laboratories or with vendors in the market, or collaborating with other external organizations. A further reason for the decentralization is that any joint R & D project involves only a small part of the technological competencies owned by a partner. Whereas the product is normally indivisible, the capacity of any unit of knowledge to generate services is theoretically infinite (Casson, 1987, p. 12). A group establishment may be invited to solve a technical problem for another firm by supplying just one unit of knowledge from its total knowledge pool. If such collaboration does not threaten or interfere with the development of the group plant or its parent, decision-making may be allowed at the establishment level. In terms of the effect of internationalization on the decision-making locus of R & D collaboration, one might jump to the conclusion that our evidence supports the arguments of Sigurdson (1990) and Westney (1990), and is against those of Porter and Fuller (1986). This is, however, not the final position taken, because R & D partnerships are probably an activity with dual characteristics. Although there are

132


indeed many collaborative R & D agreements which can be regarded as strategic alliances, joint R & D can also be an operational option for solving problems or collectively presenting bids for business (e.g. in the electrical equipment and defence-related instrument industries). Such practices are not so much related to decisions determining what types of business the company as a whole should be in, which is often defined as corporate strategy (Johnson and Scholes, 1993, p. 11), but are rather routine business operations. Of equal importance is the fact that the type of R & D partnerships can affect the locus of control. (Formal R & D partnerships have been classified by Pisano et al. (1988, p. 56) into joint ventures, equity participation, non-equity agreements and consortia. In addition, informal collaborative agreements could be more widespread than are formal collaborative agreements (Hakansson and Johanson, 1988).)Some forms of R & D collaboration, such as joint ventures and equity participation, possibly entail restructuring part of the companies involved. As a result, it is unlikely for this kind of collaborative decision to be taken at levels other than that of the group headquarters. On the other hand, informal agreements could simply take the form of customers or suppliers being drawn into collaboration or consulted in the course of R&D. These latter practices could become routine, which could allow decisions to be made by low-level tiers in the organization. It is, therefore, wrong to treat all kinds of R & D collaboration as parts of a homogeneous whole. The locus of corporate control for R & D collaboration can depend on the nature, purpose, form, and importance of the project in question. None the less, we at least are able to reject the view of Delapalme (1987) that R & D partnerships should only be seen as part of strategic alliances. First, as just argued, R & D collaboration does not necessarily involve high-level strategic decisions. Second, as joint R & D decisions can be decided at the establishment level in some, if not many, firms, this low-level tier generally has insufficient authority to extend the content of co-operative tasks to cover other functions such as manufacturing and marketing, and exchanging products and technologies. Moreover, collaborating establishments cannot alone decide the business scope of their corporations.

7. Policy implications The empirical results shown above allow one to throw some light on the public policies on R & D collaboration. Two issues are particularly relevant: pre-competitive vs. near-market collaboration, and the concept of 'additionality'. The term pre-competitive research has been widely used in policy statements to support R & D collaboration. It reflects a belief that firms will not jointly conduct research towards the competitive end of the R & D spectrum, as elaborated by Ouchi and Bolton (1988). Contrary to this belief, the evidence discussed above reveals that joint R & D activities mainly involve development, and even commercial production. To emphasize pre-competitive research as a condition for a project to qualify for public support ignores the fact that the process of innovation tends to be an iterative one. It is indeed observable from our evidence that joint research activities can just involve basic research and commercial production without touching on applied research and development. A possible deduction is that a joint R & D project may be initiated to solve problems arising from a downstream stage which then requires the undertaking of upstream research. One should also note that research closer to the market is more costly and commercially uncertain. As technological competition intensifies, firms may feel pushed, if not forced, to co-operate in near-market research. An additional rationale for public programmes to support merely pre-competitive research is the assumption that private firms can and should take responsibility for near-market R & D (see Garnsey and Moore, 1993, p. 77). Garnsey and Moore are uneasy with such a view and have argued that market failure can lead to an under-investment in nearmarket R & D by private firms. Added to this, 'institutional failure' may also require public intervention in near-market R&D. In a national context, institutional failure refers to the deficiencies of the financial system underlying business activity, the method of management, and the systems of education and training which undermine national innovative potentials and economic dynamism. The sharp contrast between the UK's scientific achievements and its economic performance could be a result of institutional failure (Sharp and Pavitt, 1993, pp.

S.-H. Chen/Research Policy 26 (1997) 121 135

141-142). Referring the concept to the firm level, institutional failure could result in a situation where a firm lacks the capacity or is unable to explore or capitalize on the results of upstream R & D , wherever they come from. Under the circumstance of institutional failure, public intervention in near-market R& D may be needed if the upstream research efforts made are to bear fruit. Apart from pre-competitive research, the concept of 'additionality', among others, has been used in rationalizing public support for (collaborative) R & D. In the simplest terms, additionality refers to the extent to which public support stimulates new R & D activity, as opposed to subsidizing what would have laken place anyway (Buisseret et al., 1995). Our finding that corporate control for R &D collaboration can be decentralized allows us to echo Buisseret and his colleagues' view on additionality in relation to firm structure. As mentioned above, a reason to decentralize control of R & D collaboration is the introduction of a customer-contractor relationship between corporate research laboratories and business units. As a result, corporate research laboratories are under pressure to seek both internal and external R & D funding. When corporate research laboratories are forced to align themselves with internal business units" technological needs, they may be diverted from fulfilling the chief mission of conducting longterm research. Publicly funded collaborative programmes can offset such an impact by providing corporate research laboratories with support for long-term research. Consequently, additionality is gained. On the other hand, aided by publicly funded programmes, corporate research laboratories may become self-sustaining to the extent that their research activities are detached from the needs of their business units and from corporate strategy. A danger, therefore, is that additionality may be gained at the expense of the coherence of corporate strategy and research efforts made by corporate research laboratories.

8. Conclusion In conclusion, we will summarize the main findings and theoretical implications. The paper has served two main purposes. First, it has presented

133

evidence which contradicts the view that firms will not form partnerships to undertake near-market research. Second, it has presented evidence rejecting the view that a decision to collaborate in R & D is made exclusively at corporate headquarters. It has been common to argue that firms are not interested in near-market collaboration. For Ouchi and Bolton (1988), the underlying reason is a requisite match between incentives to innovate and the nature of intellectual property rights. Specifically, near-market research (namely development), it is argued, can be effectively appropriated through internal R&D. However, the major type of research activity conducted by linked partners in the joint R & D process in our survey was found to be development. This is particularly apparent for private joint R & D projects, whereas public programmes were relatively more oriented to applied and basic research. The arguments of Ouchi and Bolton for joint R & D are therefore not supported. Although many researchers tend to regard R & D partnerships exclusively as a corporate-level task, evidence presented in the paper does not support such views. The locus of corporate control for R & D collaboration was found mainly at the individual establishment level, with the possible exception of the machine tool industry. More importantly, this pattern of decentralizing control appears to apply equally to home- and foreign-owned subsidiaries. To help interpret the findings, both industry-specific and more general factors have been discussed in detail. Given the patterns of decentralized corporate control of R & D collaboration and their underlying reasons, it is implausible to argue that all R & D collaboration must be pan of a strategic alliance. There is no denying that R & D partnerships can in many cases be driven by strategic aims, and hence viewed by academics and others as strategic alliances, but this is only a limited view of the situation. On the other hand, where corporate control of R& D collaboration has been decentralized, low-level tiers in a corporate hierarchy are unlikely to have sufficient authority to extend their co-operative ventures to cover other functions such as manufacturing and marketing, and exchanging products and technologies. More importantly, decentralization of R & D responsibilities does not lead to establishments designing corporate strategy alone. The outcome of this debate is to conclude

134


that R & D collaboration is an activity with both strategic and tactical orientations. Therefore, a more flexible view of the strategic nature of R & D collaboration needs to be acknowledged. R & D collaboration can be a strategic alliance or a part of a strategic alliance, but it may also be an operational tactic. The view that R & D partnerships are a strategic corporate decision, important as it is, has been overplayed.

Acknowledgements A earlier version of this paper was prepared for the First International Conference on Pacific Basin Business and Economics, held by the National Central University in Chungli, Taiwan, on 15-16 June 1996. In addition to the conference itself, the author has benefited from detailed comments by Alfred Thwaites, Jeremy Howells, Neil Alderman, and two anonymous referees.

References Amin, A., 1993. The regional development potential of inward investment in the less favoured regions of the European Community. Paper presented at the Conference on Cohesion and Conflict in the Single Market, Newcastle upon Tyne, 18-20 November. Buisseret, T., Cameron, H.M. and Georghiou, L., 1995. What difference does it make? Additionality in the public support of R&D in large finns. Int. J. Technol. Manage., Special Issue on the Evaluation of Research and Innovation, 10: 587-600. Casson, M., 1987. The Firm and the Market: Studies on Multinational Enterprise and the Scope of the Firm. MIT Press, Cambridge, MA. Casson, M. and Singh, S., 1993. Corporate research and development strategies: the influence of firm, industry and country factors on the decentralization of R&D. R&D Manage., 23: 91-107. Charles, D., 1989. Technopolis: a High Tech Formula for Urban Development, Working Paper 7. CURDS, University of Newcastle upon Tyne. Charles, D., 1990. The corporate dimension of research and development location. Ph.D. thesis, Department of Geography, University of Newcastle upon Tyne. Chen, S.H., 1994. Collaboration in Industrial Research and Development: its Nature, Rationale and Geography. Ph.D. thesis, CURDS, University of Newcastle upon Tyne. Chesnais, F., 1988. Technical co-operation agreements between finns. Sci. Technol. Ind. Rev., 4: 51-119. Delapalme, B., 1987. Introduction speech for presentation by Working Group 38: Cooperative R&D in Industry. In: Cooper-

ative R&D. European Industrial Research Management Association, Copenhagen. De Meyer, A., 1993. Internationalizing R&D improves a firm's technical learning. Res. Technol. Manage., 36. Dodgson, M., 1993a. Technological Collaboration in Industry. Routledge, London. Dodgson, M., 1993b. Learning, trust and interfirm technological linkages: some theoretical association. Paper presented at the 2nd ASEAT Conference, 21-23 April, UMIST, Manchester. Evans, A.W., 1973. The location of the headquarters of industrial companies. Urban Stud., 10: 387-395. Fransman, M., 1990. The Market and Beyond: Cooperation and Competition in Information Technology Development in the Japanese System. Cambridge University Press, Cambridge. Garnsey, E. and Moore, I., 1993. Pre-competitive and near-market research and development: problems for innovation policy. J. Technol. Manage., Spec. Publ. on Small Finns and Innovation, 69-83. Goddard, J. and Smith, I., 1978. Changes in corporate control in the British urban system. Environ. Plann., A, 10: 1073-1084. Hagedoorn, J., 1990. Organizational modes of inter-firm co-operation and technology transfer. Technovation, 10: 17-30. Hakanson, L., 1990. International decentralization of R&D: the organizational challenge. In: C. Bartlett, Y. Doz and G. Hedlund (Eds.), Managing the Global Finn. Routledge, London. Hakanson, L., 1993. Managing cooperative research and development: partner selection and contract design. R&D Manage., 23: 273-285. Hakansson, H. and Johanson, J., 1988. Formal and informal strategies in international industrial networks. In: F. Contractor and P. Lorange (Eds.), Co-operative Strategies in International Business. Lexington Books, Lexington, MA. House of Lords, 1990. Definitions of R&D, by Select Committee on Science and Technology. HMSO, London. Howells, J., 1990. The internationalisation of R&D and the development of global research networks. Reg. Stud., 24:495-512. Johnson, G. and Scholes, K., 1993. Exploring Corporate Strategy: Text and Cases, 3rd edn. Prentice-Hall, Hemel Hempstead, UK. Jorde, T. and Teece, D., 1990. Innovation and cooperation: implications for competition and antitrust. J. Econ. Perspect., 4: 75 -96. Li, F., 1993. The IT revolution and the geography of corporate restructuring. Ph.D. thesis, CURDS, University of Newcastle upon Tyne. Massey, D., Quintas, P. and Wield, D., 1992. High-Tech Fantasies: Science Parks in Society, Science and Space. Routledge, London. Mayes, D.G., 1991. The machine tool industry. In: D.G. Mayes (Ed.), The European Challenge: Industry's Response to the 1992 Programme. Harvester Wheatsheat, Hemel Hempstead, UK. Newfarmer, R., 1980. Transnational Conglomerates and the Economics of Dependent Development. JAI Press, Greenwich, CT. Norris, W., 1985. Cooperative R&D: a regional strategy. Issues Sci. Technol., (winter): 92-102.

S.-H. Chen /Research Policy 26 (1997) 121-135 OECD, 1981. Frascati Manual 1980: The Measurement of Scientific and Technical Activities: Proposed Standard Practice for Surveys of Research and Experimental Development. Organization for Economic Co-operation and Development, Paris. Ouchi, W. and Bolton, M.K., 1988. The logic of joint research and development. Calif. Manage. Rev., (spring): 9-33. Pisano, G., Russo, M.V. and Teece, D., 1988. Joint ventures and collaborative agreements in the telecommunications equipment industry. In: D. Mowery (Ed.), International Collaborative Ventures in the U.S. Manufacturing. Ballinger, Cambridge, MA. Porter, M., 1986. Competition in global industries: a conceptual framework. In: M. Porter (Ed.), Competition in Global Industries. Harvard Business School Press, Boston, MA. Porter, M. and Fuller, M., 1986. Coalitions and global strategy. In: M. Porter (Ed.), Competition in Global Industries. Harvard Business School Press, Boston, MA. Rendeiro, J., 1985. Instrumentation. In: L. Soete (Ed.), Technological Trends and Employment: 3 Electronics and Communications. Gower, Aldershot, UK. Rendeiro, J. and Shepherd, G., 1981. Technical Change and Competitiveness in the Scientific and Industrial Instrument Industry in the EEC. Report for the Commission of the European Communities, University of Sussex, Brighton. Rothwell, R. and Dodgson, M., 1992. European technology policy revolution: convergence towards SMEs and regional technology transfer. Technovation, 12: 223-238. Sharp, M. and Pavitt, K., 1993. Technology policy in the 1990s: old trends and new realities. J. Common Market Stud., 31: 129-151. Sharp, M. and Shearman, C., 1987. European Technological Collaboration. Routledge and Kegan Paul, London. Sigurdson, J., 1990. The internationalization of R&D: an interpre-

135

tation of forces and responses. In: J. Sigurdson (Ed.), Measuring the Dynamics of Technological Change. Pinter, London. Steele, L., 1975. Innovation in Big Business. Elsevier, New York. Surrey, A.J. and Chesshire, J.H., 1972. The World Market for Electrical Power Equipment. SPRU, University of Sussex, Brighton. Teece, D., 1988. Technological change and the nature of the finn. In: G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete (Eds.), Technical Change and Economic Theory. Pinter, London. Thomas, S. and McGowan, F., 1991. Procurement in European power plant. Utilities Policy, 1: 124-133. Thwaites, A.T., Oakey, R. and Nash, P. 1981. Industrial Innovation and Regional Development. Final report to the Department of Environment, CURDS, University of Newcastle upon Tyne. Von Hippel, E., 1976. The dominant role of users in the scientific instrument innovation process. Res. Policy, 5: 212-239. Westney, E., 1990. Internal and external linkages in the MNC: the case of R&D subsidiaries in Japan. In: C. Bartlett, Y. Doz and G. Hedlund (Eds.), Managing the Global Firm. Routledge, London. Whittington, R., 1990. The changing structures of R&D: from centralization to fragmentation. In: R. Loveridge and M. Pitt (Eds.), The Strategic Management of Technological Innovation. Wiley, Chichester, UK. Young, S., 1990. Japanese move the machine tool industry up a gear. Multinat. Bus., (winter): 25-35. Young, S., Dunlop, S. and Hood, N., 1992. The strategic response of multinationals in the machine tool industry. In: P. Barrar and C. Cooper (Eds.), Managing Organisation in 1992: Strategic Responses. Routledge, London.