Innovation Systems and European Integration (ISE)
A research project funded by the Targeted Socio-Economic Research (TSER) program of the European Commission (DG XII) under the Fourth Framework Program, European Commission (Contract no. SOE1-CT95-1004, DG XII SOLS), coordinated by Professor Charles Edquist of the Systems of Innovation Research Program (SIRP) at Linköping University (Sweden).
Sub-Project 3.2.2: Public Technology Procurement as a Policy Instrument
Government Technology Procurement and Innovation Theory
Submitted to the Commission : March, 1998
Charles Edquist Leif Hommen
Department of Technology and Social Change, Linköping University S-581 83 Linköping, Sweden TEL:+ 46-(0)13-282260 FAX:+46-(0)13-284461 E-mail:[email protected]
Abstract: The paper deals with public technology procurement as an instrument for innovation and industrial policy. It relates public technology procurement to various innovation theories, concentrating on those contributing to the development of a user-producer interaction perspective on public technology procurement. This perspective is utilized to identify and discuss past problems and future prospects of public technology procurement, at a national as well as a European (EU) level. The concern with developing policy implications is reflected throughout the paper. A number of different national ‘models’ of public technology procurement are compared and contrasted, and the political and institutional problems associated with public technology procurement are explicitly and critically discussed. The EC procurement rules are also discussed in relation to the theory and practice of public technology procurement, in both the main text and a lengthy appendix. The section devoted to policy implications focuses on the EU level. The paper also serves to provide a theoretical and conceptual framework and methological guidlines for the conduct of the empirical case studies of public technology procurement which were carried out in the second phase of the ISE sub-project on public technology procurement as a policy instrument. Accordingly, certain sections are devoted to the development of relevant analytical categories -- such as, for example, identification of the stages in a process of public technology procurement. The paper begins by defining public technology procurement and discussing its relation, as one of the few ‘demand side’ instruments, to other instruments of innovation policy, which are in the majority ‘supply side’ instruments. Subsequently, a classification of various types of public technology procurement is developed. Subsequently, private sector technology procurement is discussed from a ‘demand side’ perspective, which is then applied to public technology procurement. This discussion is followed by a discussion of public technology procurement as a special case of user-producer interaction, drawing upon various theories and concepts that have contributed to the Systems of Innovation approach. The concluding parts of the paper are concerned with a developing a critical analysis of the EC procurement rules and with drawing policy implications relevant to both the EU and Member States. Key-words: Case Studies; Comparative Analysis; Demand Side; European Union; Industrial Policy; Innovation Policy; Policy Instruments; Public Technology Procurement; Systems of Innovation; User-Producer Interaction. 2
! " # $% ! " & ' ! " () " * +!!, $
$ - . ! " ' $ ) ! ' $ $/ 0+ / /- 0+ / $ ' ( 01 ! # $ 2- & $ # " 3 $ 4 $
' " 5 ) - $$ ' " 6, ! 7 $ ' - " " ") $# ' $(5 ) " ") $& ' $ " 8 $9 ' $ ) $9 ' $ $) ' ' $ '+") '$ ' ' " 6 7 ' ' " ": '4 ' ! '4 ' $% !) '3 ' ' ) / 0 /) * ' # ") - * ' # ;" ) ) 6 " ) 7 ' # $< $ ' # ' " ) ' ' # 3
' 4 ! " ;! # ' 4 +" = ""! ") # ' 4 $ 8 "! ") 3 ' 4 $ ) 3 ' 4 $ $") #* ' 4 $ ')) # ' 4 ' ) 6 7 #$ ' 4 " "! ") # ' 4 / 0 /"! ") # ' 4 $/ 0 /"! ") ## ' 4 # " ) " #4 ' & #3
! ! " #3 ! )) 4 $ !: "2(! 2 4 ' / / ) ! ) 43 ' 0 8 49 ' $ - ) & ' ' ! " &' ' ,8 &9 ' # -!) 8 3 ) )) 33
# " ! " 9$
4 ( ; 9
& % ) ! 9&
)) 5( ; *$
)) 5< $#
; '4 4
1. Introduction: A Short History of Procurement
Government Technology Procurement (GTP) occurs when a government agency places an order for a product or system which does not exist at the time, but which could (probably) be developed within a reasonable period. Additional or new technological development work is required to fulfill the demands of the buyer. This is the 'ideal type' of government technology procurement.
In contrast to 'Government Technology Procurement', regular 'Government procurement' occurs where government authorities buy ready made 'simple' products such as pens and papers -- where no R&D is involved. Only price and performance of the (existing) product is taken into consideration when the supplier is selected. No such cases are studied in the ISE procurement sub-project.
In capitalist economic systems, where markets are considered to be effective mechanisms for articulating and satisfying most economic needs or demands, the point of departure in the application of government technology procurement must be the satisfaction of genuine social needs -- in other words, specific societal needs unlikely to be met by the market. The products and systems that are developed -- and the 5 technical change that enables their provision -- as the result of government technology procurement must therefore be targeted to solve specific problems.
The procurement instrument has been used in the defense material sector in many countries. It has also proved to be an extremely potent instrument for influencing the speed and direction of innovation on the civilian side in some countries -- usually in traditional infrastructure development. In addition, it has served to enhance the competitiveness of those firms successfully meeting the demands and thereby winning the contracts. Procurement was, for example, highly instrumental in the consolidation of some firms of a Swedish origin which are currently large and international. Examples are Ericsson and ABB.1
1.1. National Models of Government Technology Procurement
The experience of Sweden and the Nordic countries with Government Technology Procurement has been, on the whole, positive. The Nordic experience has been characterised by informed public purchasing and the exercise of a fairly high level of technological capacity (or 'user competence') on the part of clients during the strategic stages of technical specification and product development. These features of GTP have been credited with playing an important role in the successful development of the telecommunications industry in the Nordic countries (Granstrand and Sigurdson 1985) . The fairly high success rate of Government Technology Procurement in promoting industrial development in the Nordic world has meant that GTP has been widely accepted by the general public, by a large segment of industry, by trade unions and by most political parties.
At least in Sweden, the practice has become institutionalized to such an extent that it may be possible to refer to a 'Swedish model' of GTP. A peculiarity of this putative model is the common occurrence of close relationships between government agencies and major firms: "This has, to a large extent, taken the form of long -term intimate collaboration on joint development between large manufacturers and large customers -- a type of industrial relationship for which the new term 'development pair' has been coined" (Fridlund 1993: 4) . This 'Swedish model' of GTP, if it may be termed that, is perhaps an extreme case of a more "generally accepted" historical tendency in Europe for government selection, through the advantage conferred to innovators by the
1 6 concentration of purchases, "to favor the emergence of national champions" (Dalpé 1994) .
The 'Swedish model' however, is far from being the "one best way" of Government Technology Procurement. The 'Japanese model' is probably much more well known, though possibly less well understood outside of its country of origin. A key feature of the Japanese approach to GTP is that it relies far less on direct purchasing of new technology than on arranging technology procurement between private economic actors.2 One of the most influential accounts of Japan's economic development from the standpoint of innovation theory has drawn attention to how "The not-so-invisible guiding hand of MITI shaped the long term pattern of stuctural change ... largely ... on the basis of judgements about the future direction of technical change and the relative importance of various technologies" (Freeman 1988: 331) . Analysts working with (explicit) Systems of Innovation perspectives have for some time been interested in the "macro-entrepreneurial role" played by Japan's Ministry of International Trade and Industry (MITI) in "organizing clusters of firms and technologies into development blocs" (Carlsson and Stankiewicz 1995: 43) . However, their investigations have shown that the form of industrial organization promoted by MITI's procurement activities is quite different from that described by the term "development pair", and is far better characterised by the term "network". The 'network' aspects of the Japanese model, together with MITI's unique blend of supply- and demand-side steering mechanisms and techniques, have also been stressed by a number of authors whose work predates the Systems of Innovation approach (Allen 1981; Vogel 1980) .
A third prominent model of Government Technology Procurement is the US model, which emerged in the developmental context of the United States postwar technology policy. Certain features of this model are well known -- in particular, the central role of the US defense program. The extent to which defense-related spending and defense- related procurement historically dominated US federal expenditures on R&D has been widely publicized (Mowery and Rosenberg 1993) . Similarly, the degree to which defense contracting was responsible for the origin and subsequent development of several major commercial industries in the US -- for example, aeronautics (Mowery and Rosenberg 1982) , semiconductors (Levin 1982) , and computers (Flamm 1987) -- has been documented in numerous studies. The protectionism of US procurement policies under the "Buy American Act" is also a widely known feature of the US model of GTP (Goodman and Saunders 1985) .
2 ! ! " # $ % 7
There are, however, other equally important though less frequently noted elements of the US model -- including the influence of anti-trust policy and the commercial orientation of much US procurement for defense purposes. It has been observed, for example, that in procurement supporting the development of computers and semi- conductors, "the US military and intelligence communities pursued a very different path from that followed by Western European governments" (Mowery 1995a: 7) . In particular, they awarded contracts not only to established firms but also to new ones, and supported projects undertaken by firms that were mainly interested in commercial markets (Flamm 1988: 134) . They also sought to disseminate technical information on both technologies across a broad industrial community, rather than confining it to a 'military enclave' (Katz and Phillips 1982) . And, they made sure of having more than one supplier for a given procurement through a policy of 'second sourcing', which "served both to diffuse knowledge of advanced technology amongst a range of firms, and to facilitate the entry of new firms yo the market" (Geroski 1990: 188) . These practices, which led to substantial "spillovers" from military procurement of defense technology into civilian applications -- and which also fostered the development of competition in procurement contracts -- "appear to have been motivated by policymakers' belief that in order to create a viable military supplier community, a larger commercial industry was needed" (Mowery 1995a: 7) .
The various national models of GTP discussed above are not static. The US model has been in transition since the 1980s, and the Swedish model is also changing. Both have been affected by intensified international competition in sectors that were formerly national 'strongholds', the rise of liberal or neo-liberal orientations in economic policy and, more recently, the end of the cold war. In the US, in particular, this has led to an ambiguous "new direction" in technology policy (Ham and Mowery 1995) , motivated in part by extensive controversy on the effectiveness of defense procurement practices (Kanz 1993) . In the case of Sweden, another important factor has been integration with the European Union. Throughout the European Union, the EC now imposes a supranational regulatory regime on public procurement, following a 1985 "revolution in the Commision's public procurement policy" and the subsequent passage of legislation and other measures related to the consolidation in 1992 of a single European "Internal Market" (Martin 1996: 16) .
1.2 Failures and Problems in Government Technology Procurement
In both the US and Europe, the debate on public procurement, including technology procurement, has focused attention on failures in Government Technology Procurement. 8
In the US, at least, such a debate has been fuelled by the publication of works documenting notable failures in GTP (e.g., Cohen and Noll 1991) . To name but one fairly recent US example, the attempt to develop civilian uses for an 'on the shelf' military technology for the generation of nuclear power was spectacularly unsuccessful. Due to premature commitment to a specific commercial application, the liquid metal fast breeder reactor (LMFBR) project at Clinch River, Tennessee, was terminated in 1983 in view of the complete lack of commercial consequences after the expenditure of more than five billion US dollars (Nelson 1984) . Europe, too, has had its failures. In Sweden, an often cited procurement failure is the Mariviken nuclear power reactor contracted by Statens Vattenfallsverk (Granstrand and Sigurdson 1985: 10) . The British-French Concorde programme is widely acknowledged to have resulted in a kind of technological success -- but not a commercial one (Eads and Nelson 1971) .
Much of the recent public debate on procurement, including but certainly not limited to the discussion of failures in GTP, has drawn attention to the distorting effects on rational economic decision making of what may be broadly termed 'political considerations' (Burton 1983) . Risk avoidance on the part of elected officials, their tendency to favour short-term projects with fairly immediate pay-offs, and their preference for decisions that have no clear distributional impact (and hence, no electoral consequences) are three such characteristics of public sector demand that have been alleged to have adverse consequences on procurement decisions in particular (Cohen and Noll 1991) . On these grounds, the case can be made that "elected officials tend to be reluctant to make procurement decisions, which, according to innovation theory, could favor innovation" (Dalpé 1994: 74) .
These problematic aspects of behaviour on the part of procurers can be exacerbated by the effects of procurement contracts on the behaviour of potentially innovative firms. Government buyers tend to maintain the same suppliers over long periods of time, often on a sole-source basis, thus inducing them to specialize in public markets (Ponssard and Pouvourville 1982) . Firms with guaranteed access to public markets may tend to neglect acquiring new markets and commercial applications for their products (Cohendet and Lebeau 1987; Zysman 1977) . Both the buyer and the supplier may tend to maintain the same technology, in order to avoid the risks of innovation and the costs of adjusting to new relationships (Williams and Smellie 1985) . And, as recent evidence from the US suggests, when suppliers limited to a single client do innovate, the potential for transfer to other applications may be minimal, even though this is an intended goal of a procurement programme (Barfield 1994) . 9
There are thus both occasional failures and enduring problems in Government Technology Procurement. Governments may fail, just like markets. However, there is also a strong case to be made for GTP from the standpoint of the economics of innovation. Arguments for Government Technology Procurement can be derived from more general arguments for government intervention in technological development. Following Blume (1981: 9 - 10) , these can be summarized as follows:
1) Underinvestment in basic research.
2) Underinvestment in long-term and/or high risk R&D, e.g., radical innovation.
3) Underinvestment in socially desirable technologies or overinvestment in socially undesirable ones.
4) Knowledge imperfections among buyers.
5) Technological backwardness in some industries.
6) High risks for early buyers and adopters of an innovation.
7) Insufficient capacity for R&D in some industries, due to fragmentation and small firm size.
8) Large-scale projects.
9) Military requirements and/or economic security needs for domestic capability in strategically important technologies or supplies.
(c.f. Granstrand and Sigurdson 1985: 7 - 8)
Since Government Technology Procurement is a demand side policy instrument, the arguments in its favour are more specifically related to the nature of demand, rather than that of supply or investment. Of course, supply and demand are not unrelated to one another. The case for GTP can thus be made on the grounds that strong social needs or demands often correspond to normally weak rates of private return on investments in innovation (Mansfield and Rapoport 1977) . In the main, however, the most frequently cited arguments in favour of GTP are those which refer primarily to certain special characteristics of demand: strategic importance, largeness of scale, high risks, and high costs (Rothwell and Zegveld 1982) . With respect to these problems, there is a fairly large body of empirical research demonstrating that public demand can be a positive, though not exclusive factor in the promotion of technological development.3 Further, it
3 & ' ( ) 693&7 10 has been demonstrated convincingly that under certain market conditions, in sectors of significant technological content, the high concentration of public demand early in the product cycle acts as a potential catalyst for innovative activity (Faucher and Fitzgibbons 1993) .
These considerations, both for and against Government Technology Procurement, have informed the debate about government purchasing in Europe and the development of the EC rules governing public procurement. For the most part, both the debate and the formation of a new regulatory regime within the European Union have been more concerned with procurement in general than with the use of procurement as a policy instrument for the development of technological innovation. However, as shown in Appendix A, the EC rules have, for good reason, clearly recognized Government Technology Procurement as a special case.
1.3 Government Technology Procurement and European Integration
Because it is carried out by public authorities, procurement is not simply an economic but also a political phenomenon.4 It is not surprising, therefore, that the European debate on procurement has been ideologically charged to some extent. Generally speaking, two ideologies have been counterposed to one another: a 'free market' orientation, which "emphasizes the need to exclusively apply commercial criteria when awarding the contract" and an 'interventionist' orientation, which "regards public procurement as an instrument to realize social and economic objectives wider than mere efficiency in the use of public money" (Martin 1996: 41) .
Of these two approaches to procurement, the first, or 'free market', orientation has clearly been the dominant influence in the development of EC policies regarding public procurement. (See Appendix A) It finds a theoretical basis in classical 'economic efficiency' principles that regard the maximization of competition as the chief means of reducing costs, increasing public savings, and guaranteeing the 'fair and equitable' use of taxpayers' money (Ponssard and Pouvourville 1982: 6 -7) . This position has been buttressed by a large body of economic research and analysis indicating that broader competition in tendering or bidding on public contracts in Europe will not only result in public savings but also promote macro-economic growth through restructuring and
4 " * " " # ' 11 adjustment (Baldwin 1987; Finsinger 1988; McLachlan 1985; Parker 1990; Tovias 1990; Utterly and Harper 1993) .
The second, or 'interventionist', orientation in procurement has been severely limited by the EC procurement rules, though some allowances are still made for it. (See Appendix A.) Although alien to the spirit of the EC rules, it is highly consistent with the past practices of most national governments. Many of them, especially under the influence of Keynesian doctrines during the postwar period (Roll 1982) , made extensive use of procurement as an instrument of socio-economic policy, including industrial policy (Geroski 1990) .
It is possible to distinguish at least four instrumental functions of this general approach to the use of public procurement as a tool of policy (Jeanrenaud 1984) . First, procurement (though not GTP) has been used as a means of increasing global demand and stimulating economic activity, thereby creating employment (Keyzer 1968) . Second, it has been used to protect national industry against foreign competition (Goodman and Saunders 1985; McLachlan 1985) . Third, procurement has been used to improve the competitiveness of certain industrial sectors, by linking secure access to public markets to commitments on the part of national champions to invest in R&D (Jeanrenaud and Meyer 1984) . Fourth, it has been used to remedy regional disparities, so as to attain redistributory objectives (Jeanrenaud 1984) . A fifth function can also be added: that of using procurement to create jobs for marginal sections of the labour force (McCrudden 1994) .
Of these five instrumental functions of public procurement, nearly all have in effect been rejected in principle, if not entirely in practice, by EC policy. From an integration policy perspective "the Commission has followed a straightforward economic efficiency approach to the regulation of public procurement, thereby neglecting other perspectives applied at national level" (Martin 1996: 89) . However, some recognition has been afforded to the third, or "competitiveness" function. This may be largely due to the fact that this use of procurement is most consistent with the general principles and aims of the economic efficiency approach, which ultimately seeks to promote economic growth through enhancing the competitiveness of firms and nations.
A concern with the competitveness of European industry and firms based in EC member states can in this sense said to be fundamental to the current EC policy regarding public procurement. Thus, for example, consultancy research supporting the Cecchini Report, which ushered in the current EC procurement regime, gave considerable attention to the importance of improving "the rate of innovation, 12 investment and growth in the sectors enjoying the benefits of restructuring, with positive effects on their international competitiveness" (WS Atkins Management Consultants and Associates 1988: 7) . The report itself stressed viewing competition policy on an EC-wide basis, permitting both national and transnational mergers so that world-league EC firms can be built (Cecchini Report, cited in Business International Ltd. 1991: 6 - 7) . In this respect, the EC can be said to remain committed to "competitiveness" as an instrumental function of procurement, particularly in sectors where "European champions" may now be required to replace national ones (see, e.g., Hartley 1987) .
In addition to its 'competitiveness' ethic and ultimate concern with fostering economic growth, the EC procurement policy has, in practice, been required to make some concessions to national-level public interest. In case law, it has had to achieve compromises between the forces of economic cohesion and those of market integration (Snyder 1990) . And, more generally, it has had to recognize that in certain contexts, "overwhelming importance granted to the pursuit of efficiency is regarded as detrimental to wider administrative and political aims" (Parsons 1988) . Where the 'wider administrative aims' in question are those of enhancing competitiveness in the form of innovation, even though this may involve the limitation or even suspension of 'competition', the EC procurement policy is therefore largely permissive, rather than restrictive. (See Appendix A.) Even though this policy may have altered some of the basic 'rules of the game', Government Technology Procurement remains an area in which national governments and public agencies continue to enjoy considerable autonomy. Thus, the existence of various national models of GTP, some of which were referred to at the outset of this discussion, should not be considered moribund or irrelevant in the context of European integration.
1.4 Summary and Overview
In the initial part of this paper we have provided a 'brief history' or overview of Government Technology Procurement (GTP) as an instrument of innovation policy. We have given a working definition of GTP, indicating that its most appropriate use is to meet specific societal needs that are not readily met by normal market mechanisms. We have demonstrated that there is no "one best way" of Government Technology Procurement. Instead, it is possible to point to the existence of various national models of GTP, each displaying unique institutional features. We have made the point that some of these models are now in transition and that in a number of countries there has been wide debate about the 'pros' and 'cons' of past GTP policies and practices. In this 13 connection, we have discussed a number of failures and problems in Government Technology Procurement -- as well as the continuing existence of economic problems for which some form of GTP may be an appropriate 'solution strategy'. We have also discussed Government Technology Procurement in the context of European integration -- where, it appears, GTP, as opposed to other kinds of procurement, remains an important instrument of national policy.
In the remaining parts of this paper, we will discuss selected aspects of Government Technology procurement in greater depth. In Section 2, we will discuss the relation of GTP to other innovation policy instruments and identify different kinds of Government Technology Procurement. In Section 3 we will discuss differnces and similarities between private technology procurement and GTP in the context of different market structures and in relation to different kinds of practical problems. In Section 4, we will develop a Systems of Innovation perspective on GTP, concentrating not only on a synthesis of theoretical work but also on the general implications for public policy. In Section 5 we will present a brief overview of the EC procurement rules and their implications for the practice of Government Technology Procurement in member states. (This is discussed at greater length in Appendix A.) Finally, in Secion 6 we will discuss the future prospects for GTP as an innovation policy instrument.
2. Different Kinds of Government Technology Procurement
We will now proceed to discuss the relations between GTP and other innovation policy instruments. Policy can influence both the supply and demand sides of innovation. Both the supply of knowledge, competence, etc. and qualified demand are important determinants of innovation, and hence can be related to innovation policy.
2.1 Types of Innovation Policies and Policy Instruments
As a consequence, policy issues and instruments can be divided into the supply side and demand side. Examples of supply side policy relate to R&D, education, firm competences, the establishment of institutions which influence communication between firms and between firms and universities, etc. Examples of demand side policy include Government Technology Procurement, laws, regulations, standards and other institutions which may influence the development and diffusion of technologies. The 14 supply and demand side innovation policy instruments are illustrated in the four-field table shown in Figure 1. The most common supply side policies often have to do with generation of knowledge, or science and technology. Government funding of research is the most common example of such policy. However, there are other options such as emphasizing the diffusion and use of existing knowledge by encouraging cooperation and trust (Foray 1997) . One method of doing this is the Swedish system of technological attachés abroad, who report home to Sweden on the existence and use of new technologies in technologically advanced countries. Both of these policy instruments are shown in Figure 1. Another method is developing cooperative arrangements for technology transfers between universities and industries. Building such longer term relationships between firms or between firms and universities can involve both benefits and risks. The benefits have to do with transferring information about new technological opportunities and new user wants. The risks have to do with the potential for conflict between the culture of industry and that of the university: for example, the introduction of material incentives and secrecy into the academic community might eventually erode its ethical standards, impair its contribution to economic welfare, and frustrate its broader social purposes. This is one example of how government policy can stimulate product innovations through new instruments or through new use of existing instruments -- and the problems that may be involved in doing so.
Figure 1: Technology Policy Instruments Technology Policy Instruments Operating on: The Supply Side The Demand Side Technology (1) R&D (Research and (4) Commanding Development) Policy Technology Development (Government Funding of (Creation-Oriented Research) Government Technology Procurement) Technology (2) Accessing Technology (3) Implementing Technology Diffusion (Technical Attachés) (Diffusion-Oriented Government Procurement of Goods and Services) (Source: Edquist 1996a: 145) 15
An important demand side innovation policy instrument, which has historically enhanced product development very efficiently, is Government Technology Procurement. In Figure 1 we make a distinction between ’Creation-Oriented Government Technology Procurement’ and ’Diffusion-Oriented Government Procurement’. We will now discuss that distinction.
2.2 'Creation-Oriented' and 'Diffusion-Oriented' Procurement
This distinction means that we in the ISE procurement sub-project include not only 'creation-oriented' government procurement, i.e., cases where new products, processes or systems are created. Also, cases of 'diffusion-oriented' or 'absorption-oriented' government procurement are included. In such cases, the product or system procured is not new to the world but still new to the country of procurement. (Emphasis is put on the fact that the system needs adaptation to local circumstances -- not simply that it is new to the country.) Also in such cases a formal contract between the procurer and the supplier should have been written. Together, 'creation-oriented' and 'diffusion-oriented' government procurement constitute what we call an 'extended ideal type' of Government Technology Procurement. The majority of cases studied in the ISE procurement sub- project are of the first kind, although the second category is just as interesting and important. Much innovation is incremental and is here represented by the 'diffusion- oriented' category. Including such procurement increases the scope for a public procurement policy.
However, the 'diffusion-oriented' cases studied in the ISE procurement sub-project also include R&D or at least (incremental) innovation on the producer side. At a minimum, the products, processes or systems are adapted to local circumstances, or to the buyer -- which includes some amount of R&D or technical change. (This is, for example, the case for software compatibility.) The capability of making such adaptations is taken into account when the supplier is selected. Both kinds of procurement should therefore be called Government Technology Procurement (GTP).
'Government procurement' then is a term reserved for cases where government authorities buy ready made 'simple' products such as pens and papers -- where no R&D is involved. Only price and performance of the (existing) product is taken into 16 consideration when the supplier is selected. No such cases are studied in the ISE procurement sub-project.
Hence, we have three kinds of government procurement, two of which we call Government Technology Procurement. In order to distinguish these, we will call the 'creation-oriented' kind 'Development Government Procurement' (DGP). The 'diffusion- oriented' sort could be labeled 'Adaptive Government Procurement' (AGP).
2.3 Procurers as End-Users versus Procurers as Catalysts
We will also introduce a distinction between cases where the procuring agency is also the end-user of the product or system and when it is not. In the ’classical’ cases the buying agency, e.g. the electricity authority, the railway company or the PTT will use the product procured itself.
Alternatively the agency may serve as a catalyst, coordinator and technical resource to the benefit of the end-users. This latter kind of procurement has been practiced by NUTEK (The Swedish Board for Industrial and Technical Development) with energy saving objectives, and also in the Swedish attempt to develop a school computer.5
An example of the NUTEK activity in energy-saving is the procurement of new refrigerators in the early 1990’s. The requirement was that much less freon - which damages the atmosphere’s ozone layer - should be used in production and that the refrigerator’s energy use should be considerably lower than with earlier designs. A bidding contest was announced where the prize was an order of at least 500 refrigerators - which went to the company which could satisfy the demands. A design which could meet the demands was presented by the company Electrolux within a relatively short time. This example illustrates clearly that innovation policy through technology procurement can have other objectives besides economic ones; in this case the goal was environmental.
In the case discussed it was not a question of using the government’s own demand as it was in the cases of trains and telephone exchanges. NUTEK was not the end user of the products that were developed. Instead support was given to the buyers - builders and
5 +, - ." 17 administrators of apartment buildings - so that they could influence the suppliers. Here it concerned products for mass markets. Except for serving as a catalyst and technical resource, NUTEK here served the role of ’organizing’ the demand into a customer group.
The NUTEK model of Government Technology Procurement is a way of "empowering the users". This can be generalized by giving resources directly to the users. These resources could then be employed by the users to specify their own requirements and then to procure products meeting these requirements from potential suppliers.
A central role of both kinds of Government Technology Procurement is to articulate demand. Hence the state agencies serve as sophisticated customers. They take societal problems and needs as their point of departure. However, scientific and technological progress is a necessity to solve the problems and satisfy the needs being focused upon.
2.4. Procurement, Demand Articulation and Policy Coordination
Since government technology procurement takes societal problems and needs as its point of departure and, accordingly, is a demand side policy instrument, it is likely to be effective in identifying and articulating new demands. Therefore this instrument might be a particularly powerful one in an overall policy of supporting the development of new products.6 Powerful instruments are certainly needed - in a European situation of the coexistence of massive unemployment (counted in many millions) with large unsatisfied needs.
We have shown in other work that product innovation -- particularly in R&D intensive goods and services meeting new or unsatisfied consumer needs -- is an essential means of creating employment (Edquist, Hommen, and McKelvey 1997) . Government Technology Procurement could thus be a powerful instrument not only for solving specific societal problems but also for addressing the more general societal problem of high unemployment.
6 / " 18
It is also important to relate DGP (Development Government Procurement) and AGP (Adaptive Government Procurement) to other kinds of innovation policy instruments mentioned in Figure 1. For example, Government Technology Procurement can sometimes in a natural way be combined with R&D subsidies as part of a demand- induced process of technical change. The organization of research projects might be closely related to a procurement case. For example, it could be helpful in preparing for procurement deals.
Such research projects could focus on special ’needs and technology areas’ as candidates for technology procurement, e.g. ’infrastructure for information technology’, ’urban traffic systems’, ’the clean car’, etc. The research projects must normally have an interdisciplinary composition and can include behavioural scientists, economists, engineers, natural scientists and innovations researchers - adapted to what the character of the projects demand. Drawing on this competence such research projects should deal with questions about what is motivated from the needs point of view and what is socially desirable. In the research projects it should also be investigated if it is technically possible to satsify the needs by technical development within reasonable economic limits. In this way, results should be reached which can be a basis for the formulation of functional demands. It is a matter of getting the relationship between needs and production to work better in the long term.
2.5 Procurement and Standard Setting7
The relation between the innovation policy instruments of standard setting and government procurement is also a very important and interesting one. Standards can be regarded as ’institutions’ in the sense of ’things that pattern behaviour’ of organizations when they are involved in innovation processes.8 One example is the formulation of the NMT 450 standard for mobile telephony by the Nordic PTT’s, which defined a very early standard compared to other regions and countries. The technical standards were very specific and gave the supplying firms something to work against in their development work. This is one important reason why mobile telephony took off early and rapidly in the Nordic countries, that the current density of mobile phones is very high in these countries and that two of the three large global players (Ericsson and
7 0 +, 1 # " , ' 2 1 " & " " ' " 8 ' 3 4 5 . 19
Nokia) in the mobile telephony equipment industry developed in these countries.9 Because of the close relation between Government Technology Procurement and standard setting in some cases we will below briefly touch upon the large literature on theory and praxis of standards.
To begin with, it is useful to distinguish between cases where standard setting occurs in close connection with government procurement, and those where standard setting is simply a matter of regulating products and product markets. For example, there may be standards set by government for products which are not purchased by any public agencies. In this type of case, standard setting may be regarded as a purely regulatory activity with no relation to procurement. However, there is often some degree of overlap between these two activities. Even in regular government procurement of existing goods and services (i.e., those which can be purchased off the shelf) public agencies as 'buyers' will be constrained to observe whatever standards (if any) have been set by government.
Where an emerging or evolving technology is involved, either or both purchasing and standard setting by government may -- or may not -- decisively influence the technology's further development.10 There may also be cases in which these two instruments are used together, so that the setting of a new standard serves to define the technical specifications of a technologically 'new' product or system that is being purchased by a public agency. Where this close coordination occurs, it may be legitimate to refer to standard setting as an aspect of Government Technology Procurement (GTP).11
Much of the discussion in the literature concerns the intermediate kinds of standard setting -- i.e., those which are neither 'purely' regulation nor 'effectively' GTP. Of these, standardisation affecting the development of emerging or evolving technologies has attracted the most attention and is most pertinent to the present discussion. For the most part, recent debate has concerned technologies for which government is a large and influential purchaser, but for which there is also a (potentially, at least) large market in
9 6 7 " " " 8 9 7 ' ' ' " " * " 10 28 : ' " * ' " " " " " "" 77 ' " 11 " +, - ! 6 " " " # " " 6; <:= " 6 # > 20 the private sector. In the case of information technology (IT), therefore, the policy issue addressed has been how to manage government purchasing of IT, through standard setting and other means, in such a way as to "receive the most benefit" in other (private) sectors of the economy "without impeding the desired efficiency gains in the government operations themselves" (Cowan 1995: 201) .
With regard to standard setting, at least, one response to this issue has been to insist that the role of governments and public agencies in the standardisation process should be minimal. This first position is against standard setting by government. Governments and public agencies should only "reflect their needs as users, rather than ... speak for the needs of non-governmental users" -- because the latter function "can often be distorted by political objectives, as well as by the abuse of government procurement powers" (Rankine 1995: 193 - 194) . There is a second position, which accepts standard setting by government. Its perhaps more balanced response has been to recognise that even if government were to refrain from setting standards de jure its market weight would still lead to de facto standard setting. Government's status as a large user means that its purchasing decisions can often have major consequences in terms of the establishment of one technological standard or another. Therefore, rather than abdicating its responsibilities, government should make "careful" choices leading to adoption of "good" standards (Cowan 1995: 214) .
The second position, which accepts standard setting by government, finds considerable support in recent literature on the economics of innovation dealing with technological 'choices' (or competitions). There is wide consensus that markets tend to 'lock in' on one of several options in technology, due to positive feedback effects of initially influential choices (Arthur 1988; Foray 1989) . 'Bandwagons' are formed -- and they may be bandwagons for either superior or inferior technologies. Since government is often a very large user, its purchasing decisions can have decisive consequences for steering this process towards the adoption of one technology standard or another. Governments as buyers of new technology therefore face an implicit choice between setting high or low standards. Standards can be set informally, through the simple exercise of buying power, or the process can be formalized (Schmidt 1992) . The second strategy is favoured, for example, by the EC (see Appendix A).
The first position, against standard setting by government, emphasizes the usual ineffectiveness of the 'informal' strategy on the grounds that the normal tendency of governments (especially large ones) is to act in a fragmented way (Rankine 1995) . It cites examples of how governmental attempts at achieving central coordination can fail. In addition, this position is sceptical about government's abilities to set 'good' standards. 21
Rather than compensating for the tendency of markets to under-supply experimentation in the early life of a technology, a premature standardisation initiative by government can exacerbate this problem. However, this argument does not address those cases where standard setting and other regulatory activity by governments can have a beneficial effect on industrial innovation -- which tends to occur "when standards are set that challenge current practices and provide a clearer pointer to the types of user needs for which firms should attempt to cater" (Geroski 1990: 183) .
The caution against premature standard setting is, nevertheless, a valid one within limits. There are other reasons, as well, for governments to take a non-interventionist approach to standard-setting. Economic 'globalization' has created both opportunities and incentives for governments to take a more passive approach to standards -- that is, simply to adopt world standards and require domestic (or other) suppliers to observe them (OECD 1991a) . Costs -- including development costs -- will be lowered and multiple suppliers of a given technology will be forced to compete on price. And, setting a standard other than the world standard (if one exists) may disadvantage domestic producers. (This argument, however, tends to overlook the fact that someone must first establish a world standard.)
Of course, government passivity with regard to standards may do nothing to advantage domestic producers either. One way of balancing these priorities is to adopt national standards that do not conflict with world standards but rather encourage compatibility between different technological systems. This permits, and may promote, the development of so-called 'gateway' technologies (Cohendet and Llerena 1997: 236 - 238) . This solution allows for the importation of foreign technologies and the export of complementary or equivalent domestic technologies, thus resolving a potential trade dilemma.
Finally, and notwithstanding the considerations raised by non-interventionists, there may be compelling reasons for governments to attempt 'early' standardisation in certain, strategically important cases. For example, the early establishment of a 'good' national standard for an emerging technology can be crucial to developing a large domestic manufacturing base and capturing a large share of world markets for that technology (OECD 1991a: 95 ff) . Governments will have the strongest justifications for doing this where technological specialisations peculiar to their own 'national systems of innovation' are involved. In such cases, there will also be good grounds for the use of Government Technology Procurement (GTP) as an instrument for the further development of these technological specialisations. 22
2. 6 Summary
In this part of the paper, we considered different kinds of Government Technology Procurement (GTP) and how they differ from other instruments of innovation policy. We clearly distinguished GTP as a demand-side instrument of innovation policy from regular 'government procurement', in which government authorities buy ready made 'simple' products and no R&D is involved.
We next introduced a distinction between 'creation-oriented' GTP, or 'Development Government Procurement' (DGP) and 'diffusion-oriented' GTP, or 'Adaptive Government Procurement' (AGP). We also discussed alternative forms of GTP, differentiating between cases where the procuring agency is also the end-user of the product or system and those in which the procuring agency acts on behalf of end-users, as a catalyst. Subsequently we addressed ways to relate DGP (Development Government Procurement) and AGP (Adaptive Government Procurement) to other kinds of innovation policy instruments mentioned in Figure 1. In this context we emphasized that Government Technology Procurement involves the articulation of demand and is essentially concerned with product innovation, which often generates employment. GTP could thus be a powerful instrument not only for solving specific societal problems but also for addressing the more general societal problem of high unemployment.
Finally, we discussed the close relationship between Government Technology Procurement and standard setting. We argued that standard setting can be regarded as a form of GTP when it serves to define the functional specifications of a technologically 'new' product or system that is being purchased by a public agency. In other cases, standard setting may be a purely regulatory activity, or there may be some degree of coordination between standard setting and purchasing by government -- which may or may not influence the development of new technologies. We went on to review recent debates on standard setting in relation to new technologies, considering the arguments for and against an active government role in this area.
There are parallels to be drawn between issues raised in the discussion of standards and questions addressed in section 1.2, which discussed 'failures and problems' in GTP. Both discussions suggest that there are sound reasons to consider GTP in relation to private technology procurement, i.e. where two private organizations are involved in the development process. In the next part of this paper, therefore, we will deal, among other things, with similarities and differences between Government Technology Procurement 23 and private technology procurement. We will also return to this subject in later parts of the paper.12
3. Technology Procurement in Context: Aspects of the 'Demand Side'
In the preceding parts of this paper we have identified Government Technology procurement as an instrument of demand-side innovation policy and considered, in a general way, its relation to other policy instruments -- particularly, other demand-side instruments such as standard setting. We have also identified different kinds of Government Technology Procurement (GTP) and different national models of GTP. In the course of the discussion, we have given examples of problems and failures in GTP, and we have pointed to controversies surrounding the use of GTP as an instrument of technology or innovation policy. We have drawn attention to parallel debates about other demand side instruments, such as standard setting. We have also addressed the more general debate about 'regular government procurement', reiterating in this context the essential differences between GTP and regular government procurement of existing, standardized goods and services requiring no further R&D.
The point, then, of much of the discussion thus far has been to indicate that Government Technology Procurement is not the only instrument of innovation policy, nor is it even the only instrument of demand-side innovation policy. Certainly, an important implication of the definitions, distinctions and qualifications that we have made with regard to Government Technology Procurement is that GTP is not always the most appropriate means for government to employ in stimulating innovation and technological development. Moreover, even when Government Technology Procurement may be the most appropriate policy tool for this purpose, there must be further consideration given by government to the kind of GTP that will be most suitable and whether or not it should be used in combination with other policy instruments.
Recognizing that Government Technology Procurement is one viable instrument of innovation policy among others and indicating the circumstances under which GTP can prove to be particularly useful and effective does not mean either apologizing for Government Technology Procurement or conducting missionary activity on its behalf.
12 0 ' " * ' , < % $' ? @ ! AA " ! ! 24
Nevertheless, there is some danger that even the most objective and disinterested discussion of Government Technology Procurement, and demand-side innovation policy more generally, may be viewed in this light. This risk stems in large part from a predispostion on the part of both the attentive public and policy-makers to view demand-side interventions by government with great suspicion. There is thus a tendency to neglect "that government regulation often has a beneficial effect on industrial innovation, and is not, as popular discussion would have us believe, wholly inimical to progress" (Geroski 1990: 183). It is matched by similar neglect that "procurement policy is, in general, a far more efficient instrument to use in stimulating innovation than any of a wide range of frequently used R&D subsidies" (ibid.).
Demand side innovation policy and Government Technology Procurement often appear, then, to be ideologically charged topics. This seems to be especially the case for Goverment Technology Procurement, which is often associated, fairly or unfairly, with protectionism, monopoly and even corruption. Apparently, what is often most objectionable or provocative about the term 'Government Technology Procurement' is the word 'Government'. We will devote this much of this part of the paper, therefore, to a discussion of 'Technology Procurement' -- that is, generic Technology Procurement in which 'Government' is not specified as the buyer -- and how it is affected by various economic and technological factors. The factors considered include market structures, circumstances affecting the choice of procurement procedures or mechanisms, and considerations related to the technology life cycle. We will then relate the main points made in the course of the discussion to a policy perspective on the 'demand-side' that favours neither protectionism nor monopoly. Finally, we will return to the arguments for -- and against -- Government Technology Procurement.
3.1 Technology Procurement (without Government)
While technology procurement does occur within the private sector, it is not the dominant practice in many markets. In consumer markets, especially, demand-pull is seldom used since individual producers tend to have greater (more concentrated) control over these markets and more information about them than do individual users. In such markets, products may be improved or modified by producers in response to changing patterns of demand, but the role of users in exercising 'demand-pull' tends to be reactive rather than pro-active. (Producers, on the other hand, may actively engage in market research.) 25
In some other market contexts, users may take a more active role, demanding changes in products through organized 'customer initiatives'. The market for paper products is one well-known example of this sort of development, and the market for detergents is another.13 However, demand pull in markets for products which require investments is considerably more difficult to exercise and "requires an articulation of customer will" (Nilson 1994) It is therefore not surprising that the most concentrated and forceful expressions of customer will should be found in markets for products requiring the highest levels of investment -- namely industrial markets. In such settings, effective demand-pull will often take the form of technology procurement (which is the same as user-producer interaction).
3.1.1 Defining and modelling technology procurement
Although it is seldom recognized as such, technology procurement is widely practised in industrial markets. The practice can be defined as "the procurement by a buyer of products, services or systems, which at the time being are not available on the market and for which some element of technical development is needed" (Granstrand 1984: 9) .14
Procurement contracting may be accomplished through several different buying procedures, including closed (or sealed bid) tendering, negotiated tendering, and direct contracting with no tenders. There are a number of identifiable steps in the procurement process, though they may not always follow the same order. There are also numerous forms of technology procurement. Variation in procedures and contracts is affected by factors such as volume and frequency of orders, complexity and pace of technological development, the number and relations of buyers and sellers, and their respective competences.
There are some well established general models of buying behaviour that are applicable to the practice of technology procurement (e.g., Choffray and Lilien 1978; Robinson,
13 ' 6
Faris, and Wind 1967; Sheth 1973; Webster and Wind 1972) . Technology procurement departs from these models mainly by adding some further stages and sequences of contingent decision making. It is 'closer' in some ways, therefore, to the interactions between buyers and sellers that have been modelled for systems and plant selling (Bergström 1980; Mattson 1978) .
Granstrand (1984) uses a 'transaction costs' framework to identify technology procurement as a special form of buyer-seller interaction. He begins with two extreme cases of involvement -- a 'customer-active' paradigm in which the buyer is fully integrated "backwards" into R&D and production for internal use and a 'manufacturer- active' paradigm in which this form of integration does not exist. These two polar cases can also be called the 'fully integrated organization' and the 'market organization'. Between them, it is possible to identify a third, intermediate form corresponding to the relations involved in technology procurement -- "a (buyer/seller) cooperative paradigm or, alternatively expressed, a quasi-integrated organizational form" (Granstrand 1984: 19) .15
In an analytic comparison of technology procurement with ordinary purchase contracting in a market organization, Granstrand finds that "the attractive feature of technology procurement is its potential to smooth peaks in the perception of risk, essentially by shifting some part of the risks from seller to buyer by affecting uncertainties and sharing the cash flows" (Granstrand 1984: 22) . For the producer, commercial uncertainties are greatly reduced if specifications can be met. Technical uncertainties are reduced for the buyer by realism in technical specifications. While the buyer's economic uncertainties are actually increased to some extent, they are compensated for by an overall decrease in the level of total uncertainty and risk resulting in reduced transaction costs.
In a similar analytic comparison of technology procurement with procurement within a vertically integrated organization, Granstrand finds it more difficult to conduct an analysis of uncertainty reduction but accomplishes this by 'splitting' the vertically integrated organization into buyer and seller components. This results in a reduction of total risk-bearing ability without any necessary improvement in terms of efficiency. The implication of this analysis is that "the disadvantages of a large organization (internal procurement bias, organizational persistence and loss of entrepreneurial incentives a. o.)
15 E " 3 " 7 7 " "8 6 > 93'?@ 9&#$*#0$*&7 27 are substantial in innovative work, thus calling for some degree of disintegration or use of market mechanisms instead of administrative ones" (Granstrand 1984: 25) .
The results of the second analytic comparison are consistent with other work in the 'transaction costs' tradition pointing to the disadvantages of large organizations with respect to innovation and the need for the introduction of greater flexibility into vertically integrated organizations, particularly at later stages in the development of an industry (Phillips 1980) .16 The results of the first analytic comparison point to the inability of "pure" markets to generate product innovations (Lundvall 1985) .17 An important implication of the combined results of the two analytic comparisons is that technology procurement is conducted most effectively within neither markets nor hierarchies. Instead, it requires a 'middle way', which might be referred to as partial integration or cooperation.18
3.1.2 Exemplifying technology procurement
Although there are few systematic empirical comparisons of technology procurement with other forms of buyer-seller interaction, there is a large literature on alternative modes of contracting and subcontracting between buyers and suppliers and their relation to competitive strategies within various industries -- for example the automotive industry (Holmes 1986). Some of these studies are particularly relevant to technology procurement (e.g., Grant and Gadde 1984) . Here, we will draw upon one such study to exemplify and illustrate the practice of technology procurement.
Hellman (1993) describes the three-year cooperation between Toyota (as customer) and Nippon Steel (as supplier) which led to the latter's production in 1983 of a new type of corrosion-resistant steel sheet. The introduction of this innovation into its production process allowed Toyota to make significant improvements in anti-corrosive protection for autobodies. It not only benefited Toyota but also strengthened the competitive position of Nippon Steel.
16 * " " ! ! 8 " 61 8 9&#?1 899 7 17 0 < % %' 7 ' " 18 " # ' "" !" 8 ! " 28
Nippon Steel remains the world's largest steel maker with a 30% domestic share in crude steel production. During the 1970s and 1980s, however, the company faced poor market conditions and undertook a major rationalization, as well as an aggressive product diversification initiative. Central to the company's strategy for product innovation was a policy and organizational structure supporting long-term cooperation agreements with key customers. For its part, Toyota was during this same period not only the number one automotive manufacturer in Japan but also a main customer of Nippon Steel and a pioneer in the use of coated steel sheet. Toyota had acquired considerable expertise and competence in this area, complementary to competences and proprietary knowledge that had been developed within Nippon steel.
Discussions between the two companies, each of which was in search of a partner for technology development and both of which had collaborated in the past, led to a proposal from Nippon Steel for cooperation with Toyota. This led, a year afterwards, to the establishment of a joint R&D group. In the ensuing year, both companies participated in refining functional specifications and conducting laboratory tests and product trials, though they followed a division of labour that reflected each firm's special expertise. Coordination between the two firms made it possible for Nippon Steel to retool its production facilities prior to Toyota's formal decision to use the new material in automotive manufacture. This resulted in "a very rapid change, compared to the time it usually takes to introduce a new material in the automotive industry" (Hellman 1993: 10) .
Although only a very simple contract governed the initial collaboration between the two companies, a more comprehensive agreement concerning exploitation of research results was concluded before large-scale use of the new product commenced. This agreement met Toyota's demands that Nippon Steel license the patent rights to other Japanese steelmakers, allowing Toyota to continue its policy of maintaining multiple suppliers of all components and materials.19 In return, the contract allowed Nippon Steel to be Toyota's sole supplier for a period of 18 months. The agreement gave Toyota, over the long term, the benefit of competitive sourcing. It also benefited Nippon Steel by stimulating the growth of the Japanese steel industry as a whole via patent licensing. "For Nippon Steel, it is better to get a smaller share of a rapidly growing market than having a monopoly position in a small and slowly growing market" (Hellman 1993: 11) .
19 ! ! " F, ( " ("' 29
One rather obvious 'lesson' to be drawn from this case is that close, long-term cooperation between a supplier and a customer in the development of a technologically new product need not preclude 'competitive' relations between them. Another is that this type of supplier-customer relationship is not uncommon in many industrial markets -- including the market for steel (Håkansson 1987) . For example, the history of the Swedish steel industry -- and, more generally, the European -- also features long-term development cooperation and stable relationships between suppliers and large customers.20 Conversely, countries whose steel industries have lacked these kinds of relationships have tended to lag behind in more advanced production and product technologies. For example, the MIT study, Made in America, attributed much of the U.S. automotive industry's poor performance in adopting advanced coating technologies to traditionally tenuous links between American steelmakers and auto producers (Dertoutzos, Lester, and Solow 1989: 103) . Thus, too strong an institutional orientation towards "pure" competition may, at times, undermine important sources of national competitive advantage.
3.2 Technology Procurement and Market Structures21
We have already noted that technology procurement does not occur in all markets, but only in certain types of markets. Here we will pursue this point further, by considering the relations between different kinds of market structures and technology procurement.
Rothwell and Zegveld (1982: 93 - 99) distinguish three basic types of market structures within which technology procurement might occur.22 These market types, which refer to different configurations of demand (rather than supply), are: monopsony, oligopsony, and polypsony. According to standard definitions, monopsony refers to "the situation when there is only a single buyer in a market", oligopsony to "a type of market in which there is a fairly high degree of concentration", and polypsony to a very diffuse market in which there are many buyers with no large shares (Bannock, Baxter, and Rees 1986) .23
20 @' * " " " 3 - 3 ' - 2G( ' " " 6 , 21 0 +, 1 # " , ' 2 1 " & " " ' " 22 " " 7 77 ' " # @' " " ! ! 23 # 77 ! ! 77 ' ' " " 30
3.2.1 Monopsony
Monopsonistic markets are, of course, those in which "demand pull" is most highly concentrated. Strong demand-pull is generally considered to be conducive to effective technology procurement as a necessary, though not sufficient, condition (Mowery and Rosenberg 1979; Parkinson 1982) . Monopsony provides for the highest possible levels of demand-pull. However "pure" monopsony seldom occurs in the real world, except in conjunction with certain "natural monopolies" -- which have all, or nearly all, been public markets in most countries. Thus, there is much evidence and analysis available for a discussion of technology procurement under monopsonistic conditions in public markets (e.g.,Dalpé 1994; Faucher and Fitzgibbons 1993) . However, relatively little is known about technology procurement under monopsonistic conditions in the private sector. This is due to its extremely rare occurrence.
Despite the paucity of evidence, it is possible to point to some empirical examples of monopsony or near-monopsony in some "high technology" industrial markets located in the private sector. Microsoft Corporation's domination of the software industry might be, or soon become, one such example. However, a much more well known and well studied historical example is the domination of the US telecommunications sector by the very large private telecommunications operator, AT&T. Porter (1990) credits AT&T with having made "large scale investments" and providing (through activities such as procurement) "stimulus and investment" contributing to the development of American infrastructure as "arguably the most advanced in the world" (Porter 1990: 297) . AT&T was in this respect an exemplar of the lead role played in technology development by privately owned (and often monopolistic) companies in "communications, power generation and transportation" (Porter 1990: 297) .
Porter later indicates, however, that a prolonged lack of domestic rivalry eventually caused AT&T and similar companies to suffer a loss of dynamism and innovation. This not only resulted in "the provision of less innovative products or services to domestic customer industries" but was also "reflected in less advanced and sophisticated needs for inputs" (Porter 1990: 664) . This situation is then contrasted by Porter with more recent developments: "Telecommunications services are a hotbed of improvement and innovation in America today after the breakup of AT&T, for example, despite some misplaced early concerns" (Porter 1990: 664) . His judgement is shared by Mowery (1995a). Mowery draws attention to the "the adoption of new technologies by mature industries, a phenomenon that has sparked productivity growth and even the appearance 31 of new products ... in these industries" (Mowery 1995a: 3) . He attributes much of the current "renaissance" of US competitiveness to "the ability of the US market to sustain a position as the home of the largest mass of 'lead users' in information technologies in particular" -- and in this connection he makes special note of recent "rapid deregulation of the domestic telecommunications system" (Mowery 1995a: 19 - 20) .
International evidence lends further support to this interpretation of the US experience. Other countries have followed the US example in deregulating and in some cases privatizing telecommunications, with positive effects on the rate of innovation among telecommunications operators and their suppliers (OECD 1996a; OECD/ICCP 1996) . Although the continued existence of public telecommunications operating monopolies in many countries has slowed the pace of innovation and its diffusion, "countries such as Japan and the United Kingdom experienced a twofold increase in employment in the mobile cellular phone service market after moving from a duopoly to a competitive market and Australia experienced a three-fold increase when moving from a monopoly to a competitive market" (OECD 1996b: 108) .24 The restructuring of PTOs has also led to increased technology procurement. For example, "specialist expertise needed for core telecommunication operations (e.g., software development) is also outsourced ... creating new jobs in and beyond the telecommunications sector, but reducing direct employment by PTOs" (OECD 1996b: 116) . Supplier industries have, as a result of such trends, experienced considerable growth in employment (Glassman 1993) .
On the whole then, the more recent evidence from the US and other telecommunications sectors indicates that, despite the advantage conferred by highly concentrated demand- pull, or 'buying power', monopsony is not the ideal market structure for effective technology procurement.25 One of the main disadvantages associated with this type of market structure is that interactions with suppliers take on a highly routine character, mirroring the interaction between monopolistic PTOs and their clientele. Product innovation is abandoned in favour of standardization and price reduction.
3.2.2 Polypsony
Polypsony is, of course, the polar opposite of monopsony. In this case there are many buyers and none have very high concentrations of buying power. Technology
24 0 " ! 8 ! " 25 " % " " " , < % % 32 procurement, as we understand the term (see subsection 3.1.1), is by definition extremely unlikely to occur within this type of market structure.26 As explained in section 3, this is because producers tend to have more control over the market than do users. The obvious example is markets for consumer goods. Occasionally there are 'customer initiatives' organized around demands for improvements to products -- the demand for 'environmentally friendly' household goods, for example -- but it is normally producers, not users, who attempt to tailor product designs to perceived changes in consumer demand.27 This is product innovation based on 'market research', not 'technology procurement' (see, for example, Rothwell 1994: 635 - 636) .
The conventional pattern in supply and demand interactions underpinning the development of manufactured goods for such markets is one of "supply-push", not "demand-pull". Manufacturers will often attempt to assess customer response through various kinds of market research. Customers may also try to seek out better products, but in many cases their search will not coincide with that of the manufacturer. Both parties may be aware that the technology could be improved, but doing this on a competitive, mass-production basis will depend on this kind of coincidence. The problem is one of 'timing'. Usually, markets will eventually act to resolve this kind of problem, but in some important cases, such as energy efficiency, situations can arise where "optima will not be sought, nor maintained, under normal market conditions" (Nilson 1994: 7 ff) .
Another problem with polypsonistic markets is that of user competence. Even if individual users were able to exercise their limited buying power for the purpose of technology procurement, how capable would they be of making enlightened choices with respect to functional specifications? Much of the existing evidence suggests that in users corresponding to those who would typify buyers in a truly polysonistic market, generally low levels of user-competence would, as a rule, severely limit this capability. Thus, even those opponents of government representation of users, who champion instead the technical wisdom of free markets, tend to hold a rather low opinion of the individual consumer as a user. One authority, for example, says of these "level 3 users" of information technology that they "have little or no interest in the technology involved apart from considerations of safety, ease of use and other ergonomic factors" (Rankine 1995: 180) .
26 ! ! 3 ' ' ! 27 ,' ' $ %' !6F . 8! " # " 33
3.2.3 Oligopsony
Given that technology procurement is both oxymoronic and a practical impossibility in the context of a truly polypsonistic market structure, and given also that monoposonistic market structures, despite their high concentration of "demand-pull", ultimately tend to inhibit product innovation, oligopsony might be the 'ideal' market structure for effective technology procurement.
In oligopsony, it will be recalled, there are a number of large buyers within the market. Though none of these large byers completely controls the market, all have fairly high levels of buying power. Technology procurement is a common practice within such markets and much of the available evidence suggests that oligopsonistic market structures are highly favourable to product innovation through technology procurement. In subsection 3.2.1 we discussed the recent transition from monopsony to oligopsony in markets for telecommunications equipment and services. We pointed out in this connection that this restructuring has greatly increased the need for more advanced and sophisticated technological inputs into telecommunications operating systems, thereby stimulating product innovation.
Oligopsony appears to be in the long-term interests of firms engaged in competition based on innovation. We have illustrated this point with the example of private technology procurement given in subsection 3.1.2. There, it was shown that a supplier and a customer firm who had realized great short term benefits from close collaboration in product innovation through technology procurement both had strong incentives to maintain an oligopsonistic relationship between buyer and seller over the longer term. This arrangement allowed the buyer to benefit from efficiencies created through competition, as well as avoiding technological 'lock in'. It also enabled the supplier firm to benefit from the longer term growth of its industry, both indirectly and directly, through patent licensing.
Oligopsonistic market structures allow one oligopsonist to function as a 'quality leader' capable of de facto standard setting that will influence the decisions of other oligopsonists. Thus, the procurement activities of a single oligopsonist may have decisive effects for others and also for the suppliers of these firms, in terms of the direction and rate of innovation in technological development (Arthur 1988; Foray 1989) . This point (and its policy implications) was discussed at length in section 2.5. Oligopsony also favours the formation of 'cooperative' relationships between firms, in 34 particular those forms of 'quasi-integration' that are most conducive to product innovation (Granstrand 1984) . (See subsection 3.1.1.)
3.3 Technology Procurement and Information Problems (Auctions) 28
We have seen that technology procurement occurs within both monopsonistic and oligopsonistic market structures. The majority of industrial markets are oligopsonistic (in contrast to public markets, which are often monopsonistic). Within such markets, however, an oligopsonist acting to procure a truly new (i.e., not yet existing) technological product or system assumes, at least temporarily, the position of a monopsonist due to the novelty of the product or system that is being procured. The monopsonistic procurer, moreover, usually faces an oligopoly of potential suppliers.
In this situation, the asymmetry of information becomes a crucial problem, one that can not be resolved by the conventional market mechanism (Akerlof 1970) . Potential suppliers have more private information about the item being procured than does the buyer. The suppliers can be expected to hide information relevant to costs and quality, thus obtaining information rents, or they can establish formal or informal cartels and cooperate against the procurer. The buyer is therefore compelled to design a pricing mechanism that will require potential suppliers to produce the item at a reasonable cost. The buyer has this possibility as a monopsonist.
The alternative designs available to the procurer can be described as auction models. Following the work of Arrow (1985) on adverse selection mechanisms, auctions can be conceived as mechanisms whose inputs are bids and whose outputs are decisions on winners and payments (McAfee and McMillan 1987) . There are a number of different design criteria that may be used in auction models: public vs. secret bidding; acceptance of highest bid or second highest bid (lowest or second lowest offers, respectively, in the case of procurement); different stages of bidding, i.e., one-step vs. recurrent bidding (as in ascending auctions). Some basic assumptions must also be made in designing auctions, as to whether bidders are: risk-neutral or risk-averse; have independent or correlated valuations of cost; and tend to bid symmetrically or asymmetrically (depending on cost structures and beliefs).
28 0 +, ; @3 ( " , ' / 2 1 ," " " ' " 3 @3! * 8 ' /* H 35
Four basic types of auction models have been identified (Vickrey 1961) . The English auction is an open-bid, first price, ascending-bid auction in which payment is equal to the winner's bid. The Dutch auction is an open-bid, first-price, descending bid auction, in which payment is equal to the auctioneer's bid that is first accepted. The First-Price, Sealed-Bid auction is a one-step aution in which each bidder submits a secret bid, the highest bid wins, and payment is equal to the bid. The Vickrey auction is a second-price sealed-bid one-step auction in which each bidder submits a secret bid, the highest bid wins, and payment is equal to the amount determined by the second-highest bid.
These various auction models may be analysed by applying to them different sets of the basic assumptions briefly referred to above. Assuming independent valuation, payment as a function of bids, symmetric, risk-neutral bidders and non-collusion, the English Auction and the Vickrey auction yield equivalent dominant equilibria (Vickrey 1961) . On average, though, the outcome of all four auctions is equal according to the revenue equivalence theorem (Milgrom 1989) . Choice of an optimal mechanism can be facilitated by the revelation principle, which dictates that there should be direct reports of bidders' valuations and that they should have sufficient incentive to reveal their true preferences (Harris and Townsend 1985) .
Under standard assumptions, any traditional auction can be chosen as long as a reserve price is set that will maximize competition (Riley and Samuelson 1981) . If the further assumption is made that the bidders are risk-averse, then the auction should be a first- price, sealed-bid auction (ibid.). Under conditions of asymmetric bidding in procurement, if domestic contractors have a lower productivity than foreign ones, they should be favoured to increase competitive pressure on the latter but they (any of the former) should not be given high odds of winning (McAfee and McMillan 1987) . If it is assumed that the beliefs of bidders are correlated, then the procurer should provide as much information as possible to make beliefs more realistic, and any other auction will be superior to the first-price sealed-bid auction (Milgrom 1989; Milgrom and Weber 1982) . The latter can be improved by paying for the procurement in the form of royalties (McAfee and McMillan 1986; McAfee and McMillan 1987) . In cases of collusive bidding or cartels, reserve prices should be increased and sealed or secret bidding should be practised (McAfee and McMillan 1987) . Alternatively, negotiations should be closed (Stigler 1964) . If in later procurements quantity and quality can be varied depending on price, the auctioneer should do this to increase competition and total surplus while minimizing procurement costs (Hansen 1988) . 36
These are the main recommendations of 'auction theory' for the practice of technology procurement. While these recommendations are based on theory they can -- and should -- be related to empirical research.29
3.4 Technology Procurement and Technology Life Cycles30
The practice of technology procurement illustrates the very close connection between two aspects of innovation that are often treated separately. On the one hand, there is the development of a given technology, in which the producers can be viewed historically as progressing through the various stages of a life-cycle. On the other hand, there is the adoption of a technology, in which the buying decisions of users can be viewed as following stages of diffusion over time. Models of the technology life cycle tend to represent these two processes as mirror images of one another. They also tend to suggest that the only significant procurement decisions in the development of a technology are those made by first or early adopters, in the initial stages of the technology life cycle. However, a closer look at recent theory on technological evolution indicates that significant procurement decisions can also be made in the later stages of a technology's diffusion. There is thus in private technology procurement a distinction to be made which is equivalent to that we drew in section 2.2 between 'creation-oriented' and 'diffusion-oriented' (government) technology procurement.
3.4.1 The conventional model
The conventional model of the technology life-cycle, as developed by Utterback and Abernathy (1975) is well known. It describes how, for a given industry, the type of innovation (product vs. process), the location of innovation, and the barriers to innovation change over time. It also relates these changes to an account of changing relations between and among sellers and buyers (or producers and users). The model begins with a 'fluid' phase of technology-competition, in which there is a high level of product innovation and a low level of process innovation. A fairly large population of producing firms within the industry, each with a distinctive product concept, is in competition to capture sufficiently large shares of an emerging market. In addition to technological uncertainties, therefore, there are also market uncertainties, constituting a
29 & * ' ' 1 /* / , % % 30 0 +, 9 1 " & , ' F " / " " ' " 37 major problem for product developers. Thus, on one hand, "(t)he main source, or stimulating factor, of innovation in this phase is insight into user needs" (Hidefjäll 1997: 20) . On the other hand, "many potential customers will not be able to perceive the relevance and value of the innovation" (ibid.).
In a subsequent, 'transitory' phase of the technology life cycle, the technology- competition is resolved in favour of one or another 'dominant design, which is then allowed to mature as a product-concept (Utterback 1994: ch. 2) . In effect, this is accomplished through the procurement decisions of influential early adopters. Subsequently, market uncertainty decreases for the (one or a few) winner(s) of the technology-competition and greater attention can be given to investment in product development and production technology. The profile of process innovation is raised dramatically, while that of product innovation enters into a decline, becoming increasingly minor in character. Eventually, there is a shift to a 'specific' phase in which the industry structure becomes highly concentrated and both product and process innovation decline. Here, "the primary stimulus of innovation comes from the need to reduce costs and improve quality in production"; there is "little room for major innovations, which may disrupt the smooth and efficient production machinery or make given investments appear as sunk costs" (Hidefjäll 1997: 21) .
In broad outline, the technology-life cycle model presents an account of technology development in which being 'first' matters -- especially at the point where a dominant design emerges. It is the producing firm(s) associated with the product-concept winning the technology-competition and the influential early adopter(s) whose procurement decision(s) first leads to the establishment of a dominant design who gain the greatest advantages. For the former, the market "takes off" and they, rather than imitators, are in the best position to profit from market growth by making the transition to standardized, high-volume production (Utterback and Suarez 1993) . The latter are the first to master the "higher order concepts" that govern subsequent problem solving with the technology (Clark 1985: 245) ; they can therefore realize greater benefits from the technology than the laggards for whom they have set the agenda. In a passage well worth quoting at length, Silverberg (Silverberg 1990a) presents this as a 'unidirectional' Schumpeterian perspective in which the dominant economic factors and the key actors may change over time, but not the technology:
The technology itself ... remains more or less the same (once it has been invented) -- it is simply passed along this pipeline (perhaps undergoing in this process some slight modifications and adaptations) to reveal its economic potential and be exploited until its innovative strength has been exhausted like a squeezed-out lemon. The diffusion literature has naturally 38
embraced an analagous perspective: an innovation arrives at time zero as a consummated creation, like Venus from Zeus's brow. The problem is then to explain why it takes so damn long for it to be completely adopted by its (a priori clearly defined population of potential adopters. Non-adoption is taken as self-evidently irrational, the (Schumpeterian) heroes being the inventors and innovators, the imitators being ambiguous figures obviously necessary to the process but somehow unseemly and undeserving, while a remnant remains of conservative stick-in-the-muds doomed to economic obscurity. The only questions to be examined are what determines the rate of diffusion, what distinguishes early from late adopters and to what extent forecasting (of the rates and ultimate level of saturation) is possible. Implicit in this viewpoint is the notion that the faster the better, the earlier the time- point of adoption the better, the higher the level of diffusion the better. (Silverberg 1990a)
3.4.2 Further theoretical development
There are, however, important variations on this basic story-line. The idea that technology life-cycles can be reversed and restarted, through the occurrence of either 'competence-enhancing' or 'competence-destroying' "technological discontinuities" has been introduced (Tushman and Anderson 1986) . The first type of discontinuity tends to consolidate the dominant position of established firms, while the second type poses a threat to them from new entrants to their industry. The idea that technology, though strongly influenced by 'demand-pull', retains an evolutionary dynamic somewhat autonomous from the influence of market forces has been elaborated in the related concepts of internal 'inducement mechanisms' and 'focusing devices' (Rosenberg 1976) and 'regimes' and 'natural trajectories' in the development of technologies (Nelson and Winter 1977: 57) . These elements have been brought together in Dosi's (1982) influential model of technological paradigms and technological trajectories. In this model, the 'supply-side' is initially responsible for creating a paradigm -- "the 'universe' of possible modalities" -- but the demand-side subsequently makes progressive gains in importance, by acting (through means such as procurement) to select developmental trajectories within the paradigm, according to their "economic significance" (Dosi 1982: 155 - 156) .
Dosi's (1982) formulation makes possible a more discerning perspective on the relations among and between sellers and buyers in the historical development of a given technology. According to the classic Schumpeterian narrative, both sellers and buyers of a new technology are motivated by the desire to capture temporary monopoly gains from innovation. The first adopter(s) whose buying decision to procure an innovative 39 technology leads to the establishment of one or another seller's product-concept as a dominant design will enjoy this same status within the buyer industry. The decisive influence of the demand-side, however, is exercised only at this time; after the dominant design has been established, an oligopoly of sellers comes into being that effectively controls the technology's further development. The Dosi model agrees with this, describing a shift in the source of oligopolistic gains from (at first) Schumpeterian 'rents' on innovation to (later) the benefits of static entry barriers (Dosi 1982: 157 - 158) . However, it also points to the continuing and growing influence of the demand-side in acting to select alternative developmental 'trajectories'. This makes it possible, following earlier observations of how the technical evolution of an innovation is bound up with and related to its pattern of diffusion (Rosenberg 1972; Sahal 1981) , to conceive in a systematic way of how a technology can undergo significant changes, not only because of the technology procurement decisions made by early adopters, but also by later ones.
These implications have been brought out more clearly in later applications of the Dosian perspective to the development of modelling frameworks for the analysis of technical change (Silverberg 1987; Silverberg 1990b; Silverberg, Dosi, and Orsenigo 1988) . In this work, the observation that technical change is an ongoing and uneven process leads to an evolutionary treatment of entrepreneurial technology replacement decisions based on the assumption that these require collective learning about 'paybacks' (Silverberg 1987) . This approach seeks to explain, among other things, why there are systematic differences in the 'payback periods' calculated by different entrepreneurs. To model the 'feedbacks' presumed to generate these patterns, it relies on the concept of technological trajectory (Dosi 1982; Nelson and Winter 1982; Sahal 1985) . On this basis, the so-called 'self-organisation' approach develops the idea that improvements or modifications to a technology that make it competitive within one or another market segment "could be represented as 'branchings' along an evolutionary 'tree' and allow for classification, taxonomy and rigorous definition of technologies in question" (Grübler 1989) . Thus, new 'bandwagons' can be identified within later stages of technology diffusion.
Within this framework, entrepreneurs can be seen to face dilemmas regarding the choice between earlier or later adoption (Rosenberg 1976) . Early adopters will be, essentially, the procurers of new technology, and in order to do this successfully, they must develop 'absorptive capacity' (Cohen and Levinthal 1990) . The competency or 'learning' levels of such firms must be exceptional for them to anticipate a sufficient payback to justify the adoption of a new technology (Silverberg, Dosi, and Orsenigo 1988) . Given high enough levels of skill accumulation, 'first movers' will make the greatest gains from the 40 introduction of new technology, but for intermediate levels "a second mover or imitator may be able to capture more of the benefits by letting early innovators bear an excessive share of the development costs" (Silverberg 1990a: 185) . Or, at low levels of skill accumulation, diffusion may fail altogether. Thus, given the importance of sufficiently high levels of learning as an essential precondition for any diffusion to occur, or any benefits to be realized, firms may be compelled to pool what would normally be regarded by them as proprietary knowledge in a process of "collective invention" -- i.e., "the free exchange of information about new techniques and plant designs among firms in an industry" through which "fruitful lines of technical advance (can be) identified and pursued" (Allen 1983: 2) . Such efforts can be reproduced or amplified through certain forms of public policy such as the successful Japanese system of "pre-competitive co-operation and co-ordination" (Silverberg 1990a: 189) .
3.4.3 Implications: 'diffusion-oriented' procurement
The theoretical revisions to the conventional model of the technology life cycle and the analytic framework reviewed above make it possible to speak legitimately of 'diffusion- oriented' as well as 'creation-oriented' procurement. It is to be be emphasized that the former is not simply a matter of diffusing an unchanged technology but, rather, a form of innovation in its own right -- and one that has important economic consequences. The critical importance of high levels of competence and learning required for potential adopters to become 'early' procurers of a technology in the later stages of its development and diffusion reinforces this point. In particular, learning is required to guage correctly the timing that will yield a sufficiently high payback to justify the adoption of a new technology. Once determined this framework can provide the procurer with the basis for a strong negotiating position with suppliers and an accurate assessment of the competitive advantages to be gained from early adoption.
3.5 Summing up the Demand Side: Porter
We have now reviewed several aspects of the demand side that are of significance for the practice of technology procurement. In particular, we have considered technology procurement as a form of economic organization, the impact of market structures on technology procurement, circumstances affecting the choice of procurement procedures or mechanisms, and considerations related to the technology life cycle. We will now relate the main points to an overall policy perspective on the demand side. 41
Porter (1990) has elaborated one of the most comprehensive views of the strategic importance of the demand side for innovation and technological development. In outlining its implications for public policy he does not avoid mentioning that there is an important role to be played by Government Technology Procurement (Porter 1990: 644 - 647) . However, he is more fundamentally concerned with private sector demand, including demand taking the form of technology procurement. To this end, Porter outlines a number of other (i.e., non-GTP) 'levers' that governments can use to improve the quality of domestic demand and so improve national competitiveness (Porter 1990: 647 - 653) . With one main exception (i.e., the consideration of 'foreign aid and political ties') these levers correspond roughly to those aspects of the demand side that we have considered in this part -- or, in one instance,'technical standards', the preceding part -- of the paper. Porter's levers include: 'regulation of products and processes', 'buyer industry structure', 'stimulating early or sophisticated demand' and 'technical standards'. We will briefly review his main points under these headings, relating them to preceding sections of this paper.
3.5.1 Regulation of products and processes (affecting competition)
With regard to the regulation of products and processes, Porter draws a fundamental distinction between regulation relating to standards and that relating to competition (Porter 1990: 647) . A very positive view is offered of the regulation of standards, which can "pressure firms to improve quality, upgrade technology, and provide features in important areas of customer (and social) concern" (ibid., 647).31 There is a less positive view of the regulation of competition -- unless it encourages domestic rivalry and new firms.
With respect to the regulation of competition, there is encouragement for policies supporting internationalization of a discriminating kind -- that is policies allowing firms to outsource less productive activities but discouraging the export of highly productive domestic activities. Firms should be under competitive pressure promoting "innovation at home". And, in this context it is stressed that "Developing domestic suppliers produces more sustainable advantage than relying solely on foreign suppliers" (Porter 1990: 658) . One means of accomplishing these purposes might be the institutionalization of procurement mechanisms or auctions of the kind discussed in section 3.3 -- particularly, those appropriate to conditions of asymmetric bidding. According to 'auction theory', if domestic contractors have a lower productivity than
31 0 % : < 42 foreign ones, they should be favoured to increase competitive pressure on the latter, but they should not be given high odds of winning. (See also Appendix B.)
There is also support for policies ensuring "vigorous domestic rivalry". This means strong anti-trust legislation, avoidance of policies supporting the emergence of 'national champions', and a strong policy against "horizontal mergers, alliances, and collusive behaviour" (Porter 1990: 663) . However, antitrust law should not "be a barrier to vertical collaboration between suppliers and buyers that is so integral to the innovation process" (ibid., 663). Such advice resonates well with the discussion of market structure effects on innovation in section 3.2. It also agrees with 'auction theory's' advice (in section 3.3) concerning the need to institutionalize anti-collusive procurement mechanisms -- particularly, high reserve 'prices' and sealed bidding or closed negotiations. (See also Appendix A.) At the same time, Porter's recommendations about the importance of non-exclusive forms of vertical collaboration are in keeping with the discussion in section 3.1 of technology procurement as a form of 'quasi-integration' or partial vertical integration.
3.5.2 Buyer industry structure
Regarding buyer industry structure in fields such as health care, electric power, and telecommunications, Porter (1990: 650) stresses that "regulation or state ownership is beneficial to supplier industries if it encourages an industry to act as a more sopisticated buyer with more stringent and advanced needs". He considers, however, that this is seldom the case with state-owned monopolies, and that "In most instances, private ownership and exposure to competitive pressure create the best environment leading industries to play the role of demanding and anticipatory buyers" (ibid.: 650). The high innovation profile of the privately owned U.S. health care sector is cited as a leading example here. However, elsewhere Denmark's socialized health care system is credited with having stimulated the innovativeness and competitiveness of certain supplier industries (ibid., 651). This perspective agrees well with some of the main points made in the preceding discussion of market structure in section 3.2 -- in particular, the point that oligopsony generally has a more positive effect on technology procurement leading to product innovation in supplier industries than does monopsony.
Porter's case against regulation supporting monopsony is that "it works against local suppliers if it retards innovation and introduces conservatism into purchasing decisions" (Porter 1990: 650) . We have arrived at similar conclusions in our discussion of technology procurement as a form of economic organization in subsection 3.1.1. In 43 subsection 3.1.2 we gave an empirical example of how the institutionalization of oligopsolistic arrangements between suppliers and buyers in technology procurement can create an incentive structure that countracts tendencies towards monopsonistic arrangements and acts as a stimulus for further innovation.
3.5.3 Stimulating early or sophisticated demand
In view of the critically important role played by "early or sophisticated demand" in the development of technologically innovative products, Porter (1990: 651) places considerable emphasis on the need to "reduce the risk perceived by (producer) firms that demand will fail to materialize". Assurances of future demand are required to encourage a sufficient investment in R&D and production. Technology procurement, as we have pointed out in our discussion in subsection 3.1.1 is a form of economic arrangement that has "potential to smooth peaks in the perception of risk, essentially by shifting some part of the risks from seller to buyer " (Granstrand 1984: 22) .
Of course, shifting risks from sellers to buyers still leaves the problem of risks to be borne by buyers. Porter therefore urges a "policy of providing incentives to buyers to be early purchasers of sophisticated products" -- which, he considers, "is often more beneficial to innovation and to competitive advantage than directly subsidizing firms" (op. cit., 651). Subsidizing buyers instead of producers subjects the latter to the discipline of having to meet the former's demands, thus stimulating competitive rivalry among producer firms. Buyers are also helped to upgrade, and this "encourages the mutually reinforcing process in which a nations's buyers become more sophisticated and in turn stimulate local producers" (ibid., 651).
These points about early demand are in essential agreement with our observations in section 3.4 on the technology life cycle and its implications for technology procurement. According to the conventional model of the technology life cycle, those who derive the greatest benefits from technological innovation are the firms who first win 'technology competitions' as suppliers, and the influential early adopters, whose influential procurement decisions as buyers initiate the 'bandwagon' effects leading to establishment of a dominant design. Subsequent imitators among both buyers and sellers gain less and less benefit from the technology, which does not alter, but is eventually exhausted as an economic resource. This is the 'classical' Schumpeterian account of technological evolution, and it provides the basic rationale for encouraging 'early demand' in the form of 'creation-oriented' procurement. 44
However, we have in section 3.4.2 reviewed recent theoretical and analytical work that points to the continuing influence of the demand side in later stages of a technology's development, making possible further evolutionary branchings. In such contexts, the buyer's role in technological innovation continues to be decisive, and those buyers who become procurers of this kind require particularly high levels of competence, both to determine the correct timing and to negotiate with suppliers from a position of strength. If they can (possibly through co-operative action) accomplish these things successfully, however, they can resolve the problems of risk facing innovative early adopters and reap the post-Schumperterian benefits. There is a case to be made, therefore, for industrial policy to encourage 'early demand' not only in the form of 'creation-oriented' procurement, but also in that of 'diffusion-oriented' procurement.
3.5.4 Technical standards
As noted earlier, in subsection 3.5.1, Porter (1990: 647) . advances a positive view of regulation related to standards. His case, contra the position that standards regulation is a counter-productive intrusion into competition, has two complementary 'pillars'. One is the argument that there are many standards related to social (e.g., environmental or health) concerns that firms can not be expected to set independently, yet which can become important sources of national competitive advantage since "selling poorly performing, unsafe, or environmentally damaging products is not a real route to competitiveness in sophisticated industries and industry segments" (ibid. 648). Tough standards in these and other areas will require buyer firms to be demanding customers in technology procurement. The other argument is that those standards which are most beneficial will not be laggard or anachronistic. Rather, what is to be desired are "stringent regulations that anticipate standards that will spread internationally"; such regulations "give a nation's firms a head start in developing products and services that will be valued elsewhere" (ibid. 648). To have positive effects on technology procurement, government policy in this area needs to have a strong focus on improving demand quality "by providing accurate and complete information to buyers" (Porter 1990: 652) . Standard setting also affects the rate of supplier innovation and buyer upgrading in industry depending on its procedural efficiency, since "when basic standards are set, firms then turn their attention to rapidly developing and improving products and processes to meet them" (ibid. 652).
This position agrees well with some of the main points made in our discussion of standard setting in section 2.5. There we distinguished between standard setting used as a means of defining functional specifications in Government Technology Procurement 45 and simple standard setting. The latter, we argued, can be viewed as an important demand-side policy instrument in its own right, since it will influence the buying decisions and technology procurement activities of private firms. We reviewed positions and arguments against and for 'interventionist' standard setting by government. Relating these to the Porterian perspective, it can be seen that the position against takes little account of those kinds of issues which can not be decided upon by negotiations among private economic actors (Rankine 1995) . The position for is, of course, very close to the Porterian one, and is supported by an extensive literature on the dynamics of technological competitions (Cowan 1995) .
Given the prevalence of 'bandwagons' (Arthur 1988; Foray 1989) it is important in terms of national competitive advantage that 'good' standards be set early, particularly in areas where a given country may hope to seize a strategic advantage for domestic firms by encouraging the appropriate kinds of technology development and procurement (OECD 1991a: 95 ff) . The standards most useful for this purpose will not be parochial or protectionist. And, to balance the concerns of developing 'national' industries and participating in a 'global' economy, standards favouring so-called gateway technologies' can be developed to resolve the trade dilemmas that governments may face (Cohendet and Llerena 1997: 236 - 238) .
3.6 Government Technology Procurement Revisited
Having dealt at some length in the preceding sections with 'technology procurement' sans 'government', we will now return to the discussion of our main topic in this paper - - Government Technology Procurement (or GTP). In the remaining section of this part of the paper, we will relate the preceding overview of key aspects of technology procurement to parallel aspects of Government Technology Procurement. To this end we will organize the discussion under headings corresponding to those used in earlier sections of this part (3) of the paper.
3.6.1 Organizing government technology procurement
In section 3.1 we adopted a definition of technology procurement that was highly similar to the initial definition of Government Technology Procurement we gave in section 1, at the very beginning of this paper. We then considered the nature of technology procurement as a distinctive form of economic organization. The discussion revolved around an analytic comparison of technology procurement with, on the one 46 hand, vertical integration, and, on the other hand, market organization (Granstrand 1984) . In comparison with regular market relations, the special advantage of technology procurement as a form of organization conducive to technological innovation was shown to be its potential for shifting a part of the burden of risk from sellers to buyers and achieving a general reduction of transaction costs. In comparison with vertical integration, it was shown that the main advantage of technology procurement is that it allows for greater efficiency and flexibility than full vertical integration, thus avoiding problems such as internal procurement bias, organizational persistence, and loss of entrepreneurial initiatives. These are often cited as 'faults' of GTP.
One rather obvious implication of the above analysis for Government Technology Procurement is that protectionist policies or favouritism leading to the establishment of an effective monopoly of supply. This is the functional equivalent in GTP of full vertical integration in private sector technology procurement. Such monopolies can be detrimental, over the longer term, to both technological innovation and economic efficiency. There is, nevertheless, a case to be made for 'protectionism' in Government Technology Procurement when it shelters potentially important 'infant industries' -- i.e., "provides an assured market for new products and processes, particularly early in product life-cycles when commercial possibilities are, at best, hazy" (Geroski 1990: 189) .
Protectionism in GTP is far less justified in maintaining national champions or protected national suppliers in what are essentially mature industries, on the highly questionable assumption that a single, large established firm provides the best organizational framework for the integration of any complex R&D project (Graham 1988) . Protected markets for established firms and industries erode competitive incentives for the 'champions' and provide public authorities with "no substantive basis upon which to evaluate their performance" (Geroski 1990: 194) .
In subsection 3.1.2 we considered a prominent example of private technology procurement that involved two large firms both situated in mature industries -- steel and automobile manufacturing -- in the development of an innovative technology (Hellman 1993) . In the case examined, the procurer worked extremely closely with the supplier until the technology had been successfully developed, but then insisted on entering into a contractual arrangement that would grant the supplier an effective monopoly of supply for only a limited period of time and assurances of only a part of any purchasing of the product thereafter. The contract also stipulated that the supplier should also make the technology available to other potential suppliers through patent licensing. This 47 ultimately benefited the supplier as well through market growth based on the overall expansion of their industry.
Policies such as these in Government Technology procurement could help to eliminate some well known problems. First, they would eliminate indiscriminate protectionism leading to entrenched monopolies of supply for large, established firms. This would encourage the entry of smaller, rival firms whose efficiency and capacity for successful commercialization are demonstrably greater, creating competitive conditions that would spur (capable) champions on to higher levels of technological and commercial achievement (Baldwin and Scott 1987; Cohen and Levin 1989; Kamien and Schwartz 1982; Scherer 1980) . Second, they would encourage substantial inter-firm 'spillovers' in supplier industries, resulting in industry and market growth through the exploitation of new commercial possibilities (McKinsey and Company 1988; Rothwell and Zegveld 1985: 119 - 120) . And third, they would encourage the development of more flexible organizational arrangements than the giant, vertically integrated firm in supplier industries, through such means as decentralization, specialist subcontracting and the formation of precompetitive consortia (Ergas 1987: 59; Maddock 1983: 7) .
There is, however, a caution against taking the prescription for the restoration of competition too far and resorting to 'pure' market relations. Government Technology Procurement, as we have defined it, would be extremely difficult if not impossible to accomplish under these conditions, since it would preclude government taking an active role in the innovation process as a buyer. It will be recalled that the special virtue of technology procurement as a form of economic organization is that it distributes risk more evenly between buyer and seller, facilitating innovation that would otherwise not be carried out by the sellers.
There are, of course, alternative mechanisms by which government could ease the burden of risk borne by technology sellers or suppliers. They are, for the most part, suppply-side solutions, such as R&D subsidies. While they are certainly more consistent with market relations, however, supply side solutions such as these have their own serious drawbacks. Thus, survey evidence indicates that R&D subsidies are unduly expensive (Folster 1989; Stoneman 1987: 202) and are often used to finance projects of an inferior quality that are unable to attract private sector financing (Rothwell and Zegveld 1982: ch. 6) . Moreover, the rate of return on publicly funded R&D tends to be extremely unreliable and is often exceedingly low (Stoneman 1987) . In contrast, the evidence for public R&D contracts indicates clearly that "government procurement as a whole has a positive and substantial effect on private R&D investment" (Lichtenberg 1988) . Thus, there is a strong, though qualified, economic case to be made for the 48 effectiveness of Government Technology Procurement as an instrument of innovation policy. That case is related, as we have emphasized, to the active role played by governments or public agencies as buyers. This provides, among other things, for quality control and "the expression of a clear, consistent set of needs towards which innovative efforts can be directed" (Geroski 1990: 189) .
3. 6.2 Market Structures and government technology procurement 32
In section 3.2 we made a parallel argument about market structure on the demand side in technology procurement. The discussion was organized on the basis of the well known classification of 'monopsony', 'oligopsony' and 'polypsony' as buyer market structures within which technology procurement might or might not occur. This classification was originally applied to GTP by Rothwell and Zegveld (1982).
3.6.2.1 monopsony
In our discussion of monopsony, in subsection 3.2.1, we made the point that this condition -- domination of the buyer market by a single large buyer -- seldom occurs, except in the case of certain "natural monopolies". Though these natural monopolies have usually fallen under public ownership, we were able to examine at least one important case of effective private-sector monopsony. We demonstrated that though monopsony provides one of the essential conditions for effective technology procurement -- i.e., an extremely high concentration of buying-power or 'demand-pull' -- it has the drawback of encouraging the development over time of rigid and non- innovative relations between buyers and sellers, due to the lack of competitive incentives and market opportunities for the development of new technological products or systems. This has been the essential brief for the deregulation of mature industries, such as telecommunications .
The literature on the practice of Government Technology Procurement makes similar points about the structure of so-called 'public markets'. From our brief review of some of this literature on section 1.2, some of the main concerns about monopsony in GTP seem to revolve around the tendency for monopolists (public or private) within these sectors not only to 'lock-in' to one or another technological trajectory but also to
32 0 8 " +, 1 # " , ' 2 1 " & " " 49 reproduce the same effects in the suppliers towards whom they act as monopsonists (e.g.,Kanz 1993) . The fairly common practice in government procurement of sole- source buying can thus lead to the creation of monopoly on the supply side (Ponssard and Pouvourville 1982; Williams and Smellie 1985) . And, "If there is only one supplier in a certain market there is considerable danger of an incestuous relationship between it and government, leading to a lack of competitive pressure on this firm and, quite likely, lack of technological performance" (Rothwell and Zegveld 1982) . This is the well known-problem of 'national champions' discussed above in 3.6.1. Aversion to risk-taking , a penchant for 'quick fixes', and the desire to avoid electoral backlash from 'picking winners' can exacerbate this problem (Cohen and Noll 1991) .
Notwithstanding these considerations, the high concentration of public demand has been clearly recognized as a strong potential catalyst in technological development (Dalpé 1987) . The condition of monopsony in public demand for new technology is common in large scale projects (Faucher and Fitzgibbons 1993) , which are characterized by high levels of cost and risk together with long-term schedules (Cohendet and Lebeau 1987) and also by their indivisibility (Shulman 1980) and structuralizing effects on the productive system (Faucher 1990) . A Canadian survey of such projects found technological innovation to be "most dynamic in projects with high demand concentration", the most technologically innovative projects to be those occurring "fairly early in the development of the industry", and the highest levels of innovative activity to be associated with "projects for which there existed a critical mass of technical, financial and organizational resources" (Faucher and Fitzgibbons 1993: 182) . Critical mass could be achieved through a variety of institutional arrangements, including not only complete vertical integration within single organizations but also "strategic alliances"between the public and private sector. However, the pattern in buyer's market structure was an unambiguously positive relation between innovativeness and effective monopsony, achieved through "quasi-market institutional arrangements ... organized to control the economic and technical risk associated with the projects" (Faucher and Fitzgibbons 1993: 183) .
Monopsonistic conditions in GTP, then, may be justified from the standpoint of innovation under certain circumstances, including an early stage of industry development, a high potential for innovation, and a high level of technical risk. Properly managed, they can be used to develop important sources of competitive advantage within national economies, as a comparison of the UK and Norwegian experiences with North Sea oil has indicated (Cook and Surrey 1982) . The best-known example of this remains the role played by the US Department of Defence in the development of the U.S. semiconducter industry. Key aspects of the DoD's role 50 included a combination of R&D subsidies and judicious procurement. In technology procurement, "policies of dual sourcing helped in the diffusion of semiconductor technology and the DoD's favourable attitude toward purchasing from new technology based firms accelerated NTBF formation and growth rates" (Rothwell 1994: 645) . (See also our earlier discussion in section 1.1.) Through the adoption of similar policies, public market monopsonists can avoid the creation of monopolistic or protected supplier industries in which "Well entrenched vested interests have an almost congenital proppensity to support the status quo and their often well financed attempts to maintain it can ossify industry structure and cause performance to stagnate" (Geroski 1990: 190) .
3.6.2.2 oligopsony
Oligopsony, in which there are a number of large buyers in the market, was identified in section 3.2.3 as possibly the ideal structure for technology procurement. In Government Technology Procurement, oligopsony is a fairly common condition. This is partly due to the fact that different branches of government may make different procurement decisions. A notable example of 'oligopolistic' tendencies within government is the US government's attempt to achieve complete interchangeability of all computer equipment it purchased through the creation of a standard input/output (O/I) interface (Rankine 1995: 188 - 189) . There was strong industry opposition to this policy, based on fear of losing the competitive advantage vested in each system's distinctive architecture. When the alternative solution of declaring only one of these competing interfaces to be the standard for all government agencies, the manufacturer argued that this would prevent them (the agencies) from purchasing the most advanced technologies. This view was "shared by many of the agencies" (Rankine 1995: 189) .
Oligopsony in Government Technology Procurement is a feature of many 'infrastructural' sectors where a government authority or public agency may share the market with private firms. Rothwell and Zegveld (1982: 93) point to the example of the US electric utilities, which comprise an oligopsonistic market for capital goods related to electric power generation and transmission. They then go on to discuss the TVA (Tennessee Valley Authority), whose role in purchasing goods of this nature has exemplified "how "analagous to the role of the 'price leader' in the theory of oligopoly, one of the oligopsonists can play the role of 'quality leader' (Rothwell and Zegveld 1982) .
The role of 'quality leader' can, however, be lost by government agencies due to conditions of industry maturation and globalization under which the needs of public 51 buyers begin to diverge from those of private ones. This has apparently been the case for defence spending in the electronics market, of which it was observed some ten years ago that it could no longer serve as the "engine of electronics industry growth" in the U.K. This was because "Only when Defence electronics procurement is so massive and so 'leading edge' as to create fundamental new technologies ... is there a chance that defence spending can find civic market success" (McKinsey and Company 1988: 75) .
Similarly, in the US the Department of Defense (DoD), which in the 1960s and 1970s played a leading role in the development of the US electronics industry, has in the context of industry maturity and internationalization "evolved from being a micro- electronics technology leader ... to adopting a largely technology follower position" (Kanz 1993: 71) . Due to specialization and obsolescence, DoD's current requirements are "seriously mismatched with commercial priorities", and even the establishment of a new component supplier base will involve "further isolation from mainstream micro- electronic technologies and commercial business methodologies" (ibid.). Suggested remedies include: planning based on the recognition of "long-term, probably permanent, interdependencies among components makers" within an increasingly global industry; "increasing use of commercial standards, practices, and components, wherever feasible"; and revision of the field support service and logistics system "to permit flexibility ...and ... enhance the substitutability at higher assembly levels of components" (Kanz 1993) . Through such means, DoD might be able to maintain its prevously successful strategies of 'technology leveraging' and 'force multipliers'.
3. 6.2.3 polypsony
In section 3.2.2 we pointed out that technology procurement as such is normally a practical impossibility under polypsony, the type of market structure characterized by many buyers. In section 2.3, however, we drew a distinction between cases of Government Technology Procurement where the procuring agency is also the end user and cases where the procuring agency acts as a catalyst, coordinator and resource to the benefit of other users. Some of the most successful examples of this type of GTP have been situated in highly polypsonistic consumer markets where it has been possible to identify important societal needs that were not being met by normal market mechanisms. There appears to be an emerging role for this type of Government Technology Procurement, which 'organizes' the demand of customer groups.
Shorter product life-cycles mean that producers for "segmented" markets will have to "interact with a broad and representative range of users during the product specification 52 and development processes, in order to be able to identify functional segmentation at an early stage and plan accordingly" (Rothwell 1994: 637) . Government Technology Procurement of the 'catalytic' kind is one mechanism whereby user needs which are not prominent at the current time but which might soon constitute important sources of "functional segmentation" can be articulated and brought into the process of procuct design and development. It can also be used to improve the competitiveness of industries at an early stage in their development. One example of this is Japan's creation of a quasi-governmental rental agency, the JECC, that aided the growth of a domestic market for computers by rapidly extending the range of domestic users, thereby assisting the development of the Japanese computer industry (Anchordoghy 1988) .
In section 2.3, we called the 'catalysing' type of Government Technology Procurement the 'NUTEK kind' of GTP, referring to the experience of a Swedish government agency which has successfully "empowered users" in a number of instances where products for mass markets were involved. The NUTEK experience also provides examples of the economic case for the use of this type of GTP as a policy instrument. As we pointed out in subsection 3.2.2, a need for this kind of intervention can arise in situations where a "pay-back gap" occurs within markets.
A prominent example of the 'pay-back gap' occurs in the case of energy efficiency. It may be possible to reduce energy consumption more cheaply than to increase energy production. However, the different economic perspectives of suppliers and end-users creates a situation in which "the suppliers calculate and act with a lower rate of return than the users" (Nilson 1994: 12) . This 'pay-back gap', which takes the form of a tendency within the market "to overspend on supply and to disregard the demand opportunities" (ibid.). The risk involved in taking the latter course becomes lower as the degree of supply monopoly becomes more pronounced, so suppliers' incentives to participate in the enhancement of energy efficiency might be decreased by measures such as deregulation that would emphasize their role as profit-maximizers (ibid.). For this reason, organizing the demand side is a more viable policy intervention than deregulating or privatizing the supply side. Under such conditions, the NUTEK kind of GTP is not only justified in theory but has also proven to be highly successful in practice.
3.6.3 Information problems and GTP (Auctions) 33
33 0 8 +, ; @3 ( " , ' / 2 1 ," " @3! 53
In section 3.3, we discussed auctions as procurement mechanisms. We considered the recommendations of 'auction theory' for the design of effective procedures in technology procurement. These recommendations are also valid for Government Technology Procurement. In GTP, the procurer of new technology is a monopsonist with considerable buying power, and is under strong obligations -- and constraints -- to spend wisely.
The main difficulties faced by the government or public agency as procurer are, as in all technology procurement, information problems. The asymmetry of information stems from the indeterminate character of innovative (i.e., previously unknown) products and the possibilities of strategic use of private information or collusion on the part of an oligopoly of suppliers. These conditions dictate that in Government Technology Procurement the procurer, who has the advantage of the first move, must design a fixed pricing mechanism that will lead potential suppliers to develop and deliver the product or system being procured at a competitive cost. Thus, well designed procurement rules can minimize the burden of Government Technology Procurement on taxpayers.
We will not repeat here the main recommendations of auction theory. These are summarized in section 3.3, and also at the end of Appendix B. We will however reiterate our closing comment in section 3.3, that the recommendations of 'auction theory' should be related to empirical research. One area of possible application is the analysis of laws governing private and public technology procurement.
In this paper, we do not examine the legalities of private technology procurement. We do however, include a fairly extensive account of the EC Procurement Rules in Appendix A. Interestingly, these rules, which govern public, not private procurement, appear to follow the recommendations of 'auction theory' (cited at the end of section 3.3) fairly closely, except that they make the first-price sealed bid auction a normal practice for both 'regular procurement' and 'technology procurement'. This accommodates assumptions of risk-averse bidding and also allows the already high odds of winning enjoyed by domestic firms to be counteracted to some extent by foreign competitors. In the latter respect the rules respond to 'auction theory's' advice concerning asymmetrical bidding. In cases of technology procurement, the rules make allowance for other types of auctions: the so-called restricted and closed procedures (the latter being a negotiation rather than a true auction).
* 8 ' /* H 54
These alternative procedures respond to auction theory's advice concerning cases of correlated beliefs. In both, the auctioneer is obligated to make all relevant information available to bidders. The restricted procedure allows for reserve prices to be set and secret bidding to be practised. The closed procedure, as its name suggests, allows for closed negotiations. Thus, the rules enable procurers to follow anti-collusive strategies. There is no prohibition on variation of quantity and quality in successive auctions to increase competition and minimize costs. In these respects, the EC procurement rules might be considered to approximate 'best practice' according to auction theory.
3.6.4 Technology life cycles and government technology procurement 34
In section 3.4 we discussed technology procurement in relation to theories and models of the technology life cycle. Our discussion was organized according to a distinction which we first made in section 2.2, between 'creation-oriented' and 'diffusion-oriented' technology procurement. Though the discussion in section 3.4 did not concern government as a procurer of new technology, it is certainly possible to make this application.
3.6.4.1 'creation-oriented' government technology procurement
In subsection 3.4.1 we discussed the well known conventional model of the technology life cycle (Utterback and Abernathy 1975; Utterback 1994) . According to the conventional model, the decisive influence exercised by the demand side in the development of a given technology occurs at a fairly early stage in the life-cycle. A 'bandwagon' is formed as the consequence of the influential buying decision(s) made by one or a few major adopters to procure one of a number of competing 'product- concepts'. Subsequently, product innovation declines and the focus is on refining techniques for large scale production of the product. The initial winners of the technology-competition and the early adopters who decide its outcome are presumed to reap the greatest economic rewards from subsequent development of the technology (Utterback and Suarez 1993) .
The conventional model of the technology life cycle thus presents a strong theoretical rationale for the use of Government Technology Procurement at an early stage in the development of a given technology. There is also abundant empirical evidence to
34 0 8 " +, 9 1 " & , ' F " / " 55 support this outlook. A comprehensive survey of large scale projects involving the development of new technology in one country has found that "public procurement is generally more effective as a positive factor in technological production in earlier stages of the procuct cycle when technical uncertainties are highest", while in later stages "the shift in emphasis focuses on economic risk, often resulting in less innovative activity" (Faucher and Fitzgibbons 1993: 182) . Other work has shown that Government Technology Procurement has been especially influential in the development of original or 'world first' innovations in technology: "This seems to suggest that government orders are oriented towards high performance products, quite often in an early stage of their product cycle" (Dalpé and DeBresson 1988: 11) . For these reasons, Rothwell (1994: 646) ends an extensive review of the literature on Government Technology Procurement with the unambiguous conclusion that GTP "has its greatest impact when the technology is newly emerging".
The diminishing impact of GTP in later stages of an industry's development would appear to be one of the main historical lessons of the US Department of Defense's evolution from 'technology leader' to 'technology follower' in the now mature US electronics industry (Kanz 1993) . However, the growing divergence between military and commercial technological trajectories may be as important a factor in such cases as the late developmental stage of the industry. This is what the evidence from other US industries owing their initial 'take-off' to military Government Technology Procurement also suggests (Geroski 1990: 188) . Indeed, as Rothwell (1994: 646) observes, "there has been some concern that emphasis in the SDI program on developing highly sophisticated radiation-proof semi-conductor devices might be leading to a divergence between US civilian and defence technological trajectories in this area". He goes on to conclude that GTP's "potential for stimulating the growth of infant industries seems significant in areas where evolving public and private sector technologies are not too divergent" (ibid. 647).
3.6.4.2 'diffusion-oriented' government technology procurement
These considerations bring us to the case to be made for 'diffusion-oriented' Government Technology Procurement. In subsection 3.4.2 we considered later theoretical revisions to the conventional model of the technology life cycle -- in particular, Dosi's (1982) formulation of 'technological paradigms' and 'technological trajectories' and Silverberg's (1990a) 'self-organisation' approach to the analysis of diffusion processes. The Dosi formulation indicates that the demand side can have a continuing influence on the development of technologies, giving rise in later stages of 56 diffusion to new evolutionary 'branchings'. Silverberg's approach provides techniques for modelling such processes and emphasizes the need for especially high levels of 'collective learning' about pay-backs to resolve the entrepreneurial dilemma of choice between early and late adoption. Early adoption can be highly beneficial to the entrepreneur and essential to broader processes of economic growth and development, but where it involves genuine technological innovation it requires exceptionally high levels of 'absorptive capacity' and the ability to bargain with suppliers from a position of strength. Herein lies a strong justification for what we have called 'diffusion-oriented' Government Technology Procurement.
Less-favoured countries in Europe (and elsewhere) are generally interested in the early procurement of technologically complex products and systems that can improve their infrastructure and contribute to economic growth and the enhancement of social conditions. Particularly within the EU, transfer payments make it possible for them to make such purchases at a relatively early stage in the development and diffusion of these products or systems. This relatively early adoption of diffusing technologies, together with the inevitable requirements for modification and adaptation to local circumstances, generates strong potentials for technological product innovation in what would normally be considered the mature stage or 'specific phase' of technological development.
Situations such as this deserve close study and analysis, not only because of their high degree of interest from the stanpoint of theory and research on innovation, but also because of their strong relevance for public policy. As the 'self-organisation' model described in section 3.4.2 suggests, countries that engage successfully in 'diffusion- oriented' Government Technology Procurement will need to assess their technological opportunities very closely, in order to achieve the correct timing. If they can do so, however, they will stand to gain favourable prices from suppliers and a leading position among competitor nations, due to improved infrastructure. There may be other 'post- Schumpeterian' benefits as well.
3.6.5 Stimulating early or sophisticated demand through GTP
In section 3.5, we considered the main elements of Porter's (1990) policy perspective on the 'demand-side' and its importance with respect to innovation and the development of 'competitive advantage'. Throughout, we maintained a focus on the relation between various aspects of the demand side and (non-government) technology procurement. Here we will not review that entire discussion. Rather, we will focus on the one policy 57 instrument whose use is recommended by Porter that we have hitherto left undiscussed. We refer, of course, to Government Technology Procurement.
In Porter's (1990: 644 - 647) view, Government Technology Procurement is the most direct means that governments have of influencing demand conditions, but it can work either for or against national competitive advantage. Protected markets, usually associated with public or private monopolies, generally work "against" innovation and competitiveness. Among the domestic firms that benefit from such protection, "innovation and upgrading ... slow down" and "products and services diverge in quality and cost from those demanded internationally (Porter 1990: 645). We have made similar points throughout this paper -- for example, in section 1.2, section 2.5, section 3.2 and section 3.5, as well as in subsections 3.6.1, 3.6.2., and 3.6.3 of this section.
Government Technology Procurement, according to Porter (1990: 645 - 646) can be a "positive force" for national competitive advantage under several circumstances. The first is that it should "provide early demand for advanced new products or services, pushing its local suppliers into new areas" (ibid. 645). We have made this argument above, in section 3.6.4. The second condition is that "government agencies should set stringent product specifications and seek sophisticated product varieties rather than merely accept what domestic suppliers offer" (ibid. 645). We made this case in section 2.5 and returned to it in subsection 3.5.4. The third condition is that "Government specifications should be set with an eye to what will be valued in other advanced nations" (ibid. 645). This, too, was argued in section 2.5 and section 3.5.4. The fourth condition is that "Government procurement that makes innovation easier works to the benefit of a nation's industry" (ibid., 645). This is, of course, one of the main arguments of this paper. See, however, section 1.1, for an initial discussion of various national models of GTP. The fifth condition is that "Government procurement must include a strong element of competition if it is to upgrade the local industry" (ibid. 645). We have made the same argument throughout this paper -- for example, in section 1.2, section 2.5, section 3.2 and section 3.5, as well as in subsections 3.6.1, 3.6.2., and 3.6.3 of this section.
Finally, Porter (1990: 646) makes the point that defense procurement, though it represents a major and early market for some types of sophisticated goods, is often a "mixed blessing": "Not only is defense demand often distracting but the goals of a defense agency diverge from those best suited to upgrade industry" (ibid. 647). We have shown with reference to the U.S. experience that this was not the case in the earlier development of some major "sunrise industries in the U.S. (section 1. 1), but has become increasingly true of the later stages of their development (subsection 3.6.4.1). 58
3.7 Summary
In this part of the paper we have considered various aspects of the 'demand-side' and its importance for innovation and technological development -- first, in relation to technology procurement sans government and then in relation to Government Technology Procurement. The topics covered included the nature of technology procurement, the impact of market structures on technology procurement, the design of effective procurement procedures to overcome information problems in 'auctions', and theory and evidence concerning the technology life cycle. The discussion covered a wide range of theory and research on innovation and technological development. In returning to the main subject of this paper, Government Technology Procurement, we not only drew forth the implications for Government Technology Procurement but also linked this discussion with our earlier discussion, in part 2, of different kinds of Government Technology Procurement.
We hope to have shown in the course of this discussion that Government Technology Procurement is not necessarily linked with protectionism, monopoly and favouritism. We also hope to have demonstrated that, as Porter (1990) argues, Government Technology Procurement can be a "positive force" for innovation and the development of competitive advantage. We have indicated the conditions under which GTP is most likely to play this role as a powerful instrument of demand-side innovation policy. We have also indicated where this use of GTP may be inappropriate, and other policy instruments -- or none -- may be used instead.
In the next part (4) of this paper, we will elaborate an 'innovation theory' perspective on Government Technology Procurement.
4. Innovation Theories and Government Technology Procurement
We have seen that innovation policies can be classified as demand side oriented or supply side oriented and that GTP is heavily demand side biased. Similarly, theories of the innovation process can be classified as being linear or systems oriented. There are important parallels and logical connections to be drawn between these two classifications. Here, we wish to draw attention to the congruence between linear views of the innovation process and the supply-side orientation in innovation policies. 59
The so-called "linear model" of innovation has been generally accepted throughout much of the period since World War II (Kline and Rosenberg 1986: 285) . A linear view of the innovation process means that "science leads to technology and technology satisfies market needs" (Gibbons et al. 1994: 51) . In other words, the linear model conceives of research and development as 'applied science' and envisions a smooth uni- directional flow from basic scientific research to commercial applications: research leads to development, which leads to production, which leads to marketing. In this model there are no feedbacks from the several later stages of the innovation process (i.e., product development, production, and marketing) to the initial stage of research, nor are there feedbacks between any of the other stages.
The linear view is very simplistic (and, we will argue below, unrealistic). However, it has derived considerable legitimacy from being highly consistent with neoclassical economic theory's 'market failure' explanations of the need for public support of industrial R&D, both directly (through subsidies) and indirectly (through the funding of basic scientific research) (Arrow 1962a; Nelson 1959) . The problem with these theoretical formulations is that they are not very helpful in practical terms. They do not assist the development of specific technology policies because they do not indicate the amount of government intervention required, the particular fields in which it is required, or the type of intervention that is required (Edquist 1994a: 70 - 71) .
The systemic view of innovation, the theoretical formulations with which it is aligned, and some of their broad policy implications will be discussed in subsequent sections (4.1 - 4.3) of this part (4) of the paper. In what immediately follows, however, we will concentrate primarily on outlining practical problems with the linear view of innovation and how the identification of these problems led to the development of a more systemic view -- and thence to the Systems of Innovation (SI) approach(es).
Problems with the linear model of innovation have been summarized by Kline and Rosenberg (1986). The complete absence of feedback paths in the model has already been noted. According to these authors, "the shortcomings and failures" that are "part of the learning process that creates innovation" mean that in both radical and evolutionary innovation "feedbacks and trials are essential" (Kline and Rosenberg 1986: 286) . Another problem with the model is its view of research as 'applied science'. It is often the case that basic scientific research does not lead to the design of innovations. Rather, "problems that are thrown up by the processes of designing and testing new products and new processes often spawn research -- true science -- and have in some instances even given rise to new branches of mathematics" (Kline and Rosenberg 1986: 286 - 60
287) . It may also be the case that technological innovations will proceed independently of any interaction with science -- though other types of interactions might be important. These and other considerations support an alternative "chain-linked model" of the innovation process (Kline and Rosenberg 1986: 289 - 290) . This model is characterized by a central path beginning with 'design' rather than 'research' and by multiple feedback paths :
These feedback paths iterate the steps and also connect back directly from perceived market needs and users to potentials for improvement of product and service performance in the next round of design. In this sense, feedback is part of the cooperation between the product specification, product development, production processes, marketing and service components of a production line. (Kline and Rosenberg 1986: 289)
Clearly, this is a more systemic view of the innovation process, in the sense that it explicitly recognizes the potentially complex interdependencies and possibilities for multiple kinds of interactions between the various elements (or 'stages') of the innovation process.35 It is also evident from the example of the chain-linked model that a systems oriented view of innovation accords great importance to the 'demand side', rather than concentrating primarily, if not exclusively, on the supply side (as in the linear model).
In our previous discussion (in part 2) we saw that GTP is a demand-side innovation policy instrument, and that a central role of Government Technology Procurement is to articulate demand. Thus, GTP is neither understood nor explained well by standard economic theory's supply-oriented 'market failure' explanations of the need for government intervention in the innovation process, nor is it adequately represented by the associated linear model of innovation.
It is also obvious from our earlier discussion (in part 2) that GTP is a matter of interaction between users and producers (manufacturers). This interaction is not only a matter of price signals and quantities sold and bought. The interaction contains other kinds of information and also knowledge, and it can initiate learning processes. GTP is therefore not captured well by a linear view (theory) of innovation.
35 8 " ' " ' " 3 3 7 77 " * 3 ! ! ' 8 " " ! 8 ! 77 " * 3 " + ' " " 8 + ' 7 8 " " , " + ' " 3 61
For these reasons, we contend that GTP is better captured by systems oriented approaches to the understanding of innovation processes. We will therefore devote much of our effort to dealing with those kinds of approaches in the following sections.
4.1. Systems of Innovation Approaches
A logical extension of systems-oriented theories of innovation has been the recent emergence of Systems of Innovation (SI) approach(es) (Lundvall 1992; Nelson and Rosenberg 1993b; Niosi et al. 1993) . Theorizing about Systems of Innovation includes non-linearity and interdependence among its key characteristics, and it is therefore natural to emphasize the demand side policy instruments in this approach. Let us start by listing nine characteristics that the various Systems of Innovation (SI) approaches have in common.
The emergence of Systems of Innovation approaches and their characteristics have been dealt with by Edquist (Edquist 1997a: section 3) .36 Nine characteristics of the approaches were identified:
1. They place innovation and learning processes at the centre of focus -- This is based on the understanding that technological innovation is a matter of producing new knowledge or combining existing elements of knowledge in new way. It is thus, in the broadest sense, a 'learning process'.
2. They adopt a holistic and interdisciplinary perspective -- They are'holistic' in the sense that they have the ambition to encompass a wide array -- or all -- of the determinants of innovation that are important. They are 'interdisciplinary' in the sense that they allow for the inclusion not only of economic factors but also of organizational, social and political factors.
3. They employ historical perspectives -- Since processes of innovation develop over time and involve the influence of many factors and feedback processes, they are best studied in terms of the 'co-evolution' of knowledge, innovation, organizations, and institutions.
36 " " " , " + ' " ' $ " 5 + * " 1 : 9 $ 6 % 62
4. They stress the differences between systems, rather than the optimality of systems -- They make the differences between systems of innovation a main focus, rather than something to be abstracted away from. This means conducting comparisons between existing systems, rather than between real ones and an ideal or 'optimal' one.
5. They emphasize interdependence and non-linearity -- This is based on the understanding that firms almost never innovate in isolation but interact more or less closely with other organizations, through complex relations that are often characterized by reciprocity and feedback mechanisms in several loops. This interaction occurs in the context of institutions -- e.g., laws, rules, regulations, norms and cultural habits. 37 Innovations are not only determined by the elements of the systems, but also by the relations between them.
6. They encompass product technologies and organizational innovations -- This is based on the understanding that developing a differentiated concept of innovation -- i.e., one which is not solely restricted to the conventional emphasis on process innovations of a technical nature -- is necessary to comprehend the complex relations between growth, employment and innovation.
7. They emphasize the central role of institutions -- They do so in order to understand the social patterning of innovative behaviour -- i.e., its typically 'path-dependent' character -- and the role played by institutions in the sense of norms, rules, laws etc. and by organizations.
8. They are still associated with conceptual diffuseness -- Thus, further development will involve making progress from the present state of 'conceptual pluralism' to a clearer specification of 'core concepts' and their precise content -- a gradual selection process in which pluralism and ambiguity will be reduced by degrees.
9. They are conceptual frameworks, rather than formal theories -- Recognizing that the approach is not yet at that stage of development where it is capable of 'formal' (abstract) theorizing leads to an emphasis on empirically-based 'appreciative' theorizing. This
37 " 3 " 4 5 8 3 " " ! " ! ! ! E ' " " C " ' ' ' C' " " 3 C" * C 4 5B % % $ 63 theorizing is intended to capture processes of innovation, their determinants, and some of their consequences (e.g., productivity growth and employment) in a meaningful way.
As indicated by this discussion, innovations and learning are placed at the very centre of focus in the Systems of Innovation (SI) approach. We will take as our point of departure in this paper the stong emphasis placed on "learning", its linkages with evolutionary and interactive theories of innovation and the affinities of both of these theoretical influences with institutionalist perspectives. These linkages and affinities are discussed in the following section (4.2). We will then go on (in section 4.3) to discuss in more detail theories of interdependency in Systems of Innovation. This review of theories and concepts related to the Systems of Innovation (SI) approach will form the basis for the considerations in subsequent parts (5 to 7). Here, we will comment only briefly on the general relevance of the immediately following theoretical discussions (in the remainder of part 3) for the later discussion of policy.
Conceptualizing Systems of Innovation (SIs) as evolutionary systems in which institutions 'matter' and learning processes are of central importance has significant implications for the development of firm strategies and public policies (McKelvey 1997) . From this perspective, policy is essentially a question of supporting interactions in a system that identify existing technical and economic opportunities or create new ones. In other words, the degree of innovation opportunity should be the deciding criterion in allocating support for certain types of interactions and, hence for certain technologies and sectors. Moreover, the feasibility of alternative directions for innovation must also be evaluated, so that policy does not remain 'blind' and support all alternatives in an indiscriminate way. Policy-makers should develop selection criteria -- such as the impact on economic growth and employment -- while supporting the generation of novelty. Accordingly, "Making these criteria explicit in terms of the economic and technical dimensions of innovation opportunities is a start towards informed decisions and policy-making" (McKelvey 1997: 243) .
In this paper, we will consider Government Technology Procurement (GTP) as a specific type of policy instrument. Our theoretical discussion is intended to clarify how and why this instrument can be used as a mechanism for selection among alternative pathways of innovation and economic development. In later sections of the paper, we will consider some of the explicit criteria that have been, or could be, applied in the use of Government Technology Procurement (GTP) as a policy instrument
4.2. Interactive Learning, Evolutionary, and Institutional Theories 64
It has earlier been argued that the development of the Systems of Innovation (SI) approach(es) has been influenced by different theories of innovation such as interactive learning theories and evolutionary theories (Edquist 1997a: 8) . The discussion of some theoretical roots of the Systems of Innovation approach(es) there was, however, very brief and had no aspirations of being complete. We will below try to go somewhat deeper into the antecedents of the Systems of Innovation approach in one particular respect. We will, in section 4.3, discuss contributions that focus upon interdependencies between elements (organizations and institutions) which influence processes of innovation. However, we will first summarize the earlier discussion of the roots of the systems approach to innovation.
What follows (in section 4.2) is a very brief discussion of some of the theoretical roots of the Systems of Innovation (SI) approach(es). This discussion will make special reference to some of the major works representing main variants of the Systems of Innovation (SI) approach. These variants include a 'sectoral' or "technological systems" approach (Carlsson 1995) and both 'broad' (Lundvall 1992) and 'narrow' (Nelson 1993) versions of a "national systems" approach.38
As indicated above, (in section 4.1, point 9 ), the Systems of Innovation (SI) approach is not considered a formal and established theory here, but rather a conceptual framework (or a set of such frameworks) useful for the generation of hypotheses and empirical generalizations. 39 As a conceptual framework, the development of the SI approach has been influenced by different theories of innovation (such as interactive learning theories and evolutionary theories).40
Lundvall’s (1992) book is an attempt to relate the national systems of innovation approach explicitly to interactive learning theories of innovation. In it, contributing authors from Aalborg University, Denmark have placed their own previously developed innovation theories into a "national systems of innovation" conceptual framework. This theory stresses processes of learning and user-producer interaction.41 In Lundvall's words:
38 E !! !! " C C' " ! ! " C C' 5B $ % 7 $ % % 39 , 5B % " 40 The remainder of this subsection is partly based on Edquist (1997a: section 2.2) 41 " ' " 7 , ' " ' 9 $B = 65
One of our starting points is that innovation is a ubiquitous phenomenon in the modern economy. In practically all parts of the economy, and at all times, we expect to find on-going processes of learning, searching and exploring, which result in new products, new techniques, new forms of organization and new markets. In some parts of the economy, these activities might be slow, gradual and incremental, but they will still be there if we take a closer look. (Lundvall 1992: 8)
The Systems of Innovation approach is compatible with the notion that processes of innovation are, to a large extent, characterized by interactive learning. It could be argued that some kind of Systems of Innovation approach is inherent in any perspective that sees the process of innovation as interactive; the notion of interaction paves the way for a systemic approach.
Several innovation theorists have convincingly argued that the model of the isolated profit maximizing firm is an inappropriate tool for interpreting certain important aspects of the processes involved in generating and diffusing innovations. For example, many of the actors and organizations involved in R&D and processes of innovation are not primarily governed by profit seeking motivations. These actors may be governmental or private non-profit organizations such as universities and public research laboratories. Legal conditions, rules, and norms will also significantly affect an organization’s inclination and possibility to innovate. Non-profit organizations and profit-oriented ones, like firms, also interact with each other in complex ways when pursuing learning and innovation (Nelson and Winter 1977: 50 - 52) .
As an alternative to understanding technical change to be a result of seeking to maximize profits, Nelson and Winter propose that it can be understood as an evolutionary process (Nelson 1987; Nelson 1995a; Nelson and Winter 1977; Nelson and Winter 1982) . An evolutionary theory of technical change often contains the following components:
1. The point of departure is the existence and reproduction of entities like genotypes in biology or a certain set-up of technologies and organizational forms in innovation studies.
2. There are mechanisms that introduce novelties in the system, i.e., creates diversity. These mechanisms include significant random elements, but may also produce predictable novelties (e.g., purpose-oriented development work). In biology the novelties are mutations and in our context they are innovations. 66
3. There are mechanisms that select among the entities present in the system. This increases the relative importance of some and diminishes that of others. The selection process reduces diversity and the mechanisms operating may be the 'natural selection' of biology or the 'market selection' of competition as regards technical change.42 Together the selection mechanisms constitute a filtering system that functions in several stages and leads to a new set-up of, for example, technologies and organizational forms. There might also be feedback from the selection to the generation of new innovations.
Nelson writes, "Technical change clearly is an evolutionary process; the innovation generator keeps on producing entities superior to those earlier in existence, and adjustment forces work slowly" (Nelson 1987: 16). The technologies that are developed are only superior in a relative sense, not optimal in an absolute sense, and the system never reaches a state of equilibrium. Technological change is an open-ended and path dependent process where no optimal solution to a technical problem can be identified.
Although the resource bases of various national systems of innovation highly influence their patterns of innovation, technical change will likely also involve considerable randomness. In addition, the processes through which new technologies are screened, selected, and implemented take considerable time. Randomness, combined with the time-consuming nature of innovation processes, may indicate that evolutionary models of technological change are a more realistic way of grasping and understanding innovations than the models provided in neoclassical economics (Nelson 1981: 1036, 1059 - 1061) .43
Evolutionary theory then is in addition to interactive learning theory, one of the theoretical perspectives that has strongly influenced the development of the conceptual framework(s) that we have referred to as the Systems of Innovation (SI) approach(es). Moreover, there are implicit (and sometimes explicit) in both the evolutionary and interactive learning theories of technical innovation certain institutionalist perspectives that are absent from standard, or neoclassical, economic theory. We alluded to implicit institutionalist perspectives of this kind in our statement, above, that evolutionary theory tends to reject the isolated profit maximizing firm as an appropriate tool for understanding some important aspects of innovation processes. Institutions 'matter' to this school of thought because it bases explanations of evolutionary patterns of change
42 + ' " 78 ' ' 8 "' ' 43 " ' " 67 on the decisions and actions of agents in relation to institutions and organizations (Hodgson 1991) .
Some authors working within the tradition of interactive learning theory are explicit about the importance of institutionalist perspectives. They draw attention to how innovation processes, like almost all learning processes, are "influenced ... by the institutional set-up of the economy" (Johnson 1992: 23) . This analytical focus stands in contrast to that of neoclassical economics. It is informed by the view that, in concentrating its attention on the issue of responsiveness to price signals, mainstream economic theory has neglected the analysis of innovation as a process guided to a large extent by "responses to other signals and responses transmitted through other forms of social interaction than market exchange" (Nielsen 1991: 5 - 6) .
Despite the importance of these theoretical influences, they are not always overtly recognized in the main formulations of Systems of Innovation (SI) approaches. This is especially true in the case of evolutionary theory. For example, there is no mention of evolutionary theory in Nelson and Rosenberg (Nelson and Rosenberg 1993a) , and their version of the national systems of innovation approach is therefore not explicitly based in evolutionary theory. However, in his 1988 article on the US innovation system, Nelson explicitly argues that "in capitalist countries, technical change is set up as an evolutionary process" (Nelson 1988: 313) . In Nelson’s (1995a) detailed survey of evolutionary theorizing there are no indications that he has abandoned the evolutionary perspective that he helped to establish (Nelson and Winter 1977, 1982). Therefore, it seems perfectly safe to assume that the Nelson (and Rosenberg) version of the Systems of Innovation approach is implicitly based on an evolutionary theory of innovation.
The Carlsson approach is, however, explicitly based on an evolutionary perspective: "We have chosen an evolutionary approach because of its ability to bring within a single conceptual framework the institutional/organizational as well as the cognitive/cultural aspects of social and economic change" (Carlsson and Stankiewicz 1995: 23) . Hence they have a fairly long discussion of the role of creation of diversity and of selection mechanisms.
In contrast, Lundvall does not refer to evolutionary theory in his introduction (Lundvall 1992). However, in chapter 4 of Lundvall (1992) Esben Sloth Andersen explicitly discusses an evolutionary framework as well as the evolutionary foundations of learning-by-doing (Andersen 1992) . We have already shown that theories of interactive learning lie behind the Aalborg version of the national systems of innovation approach. We have also stressed the connection between theories of interactive learning and 68 institutionalist perspectives (see, e.g., Lundvall 1992: 8 - 13) .) And, we have indicated that evolutionary theories have a similar affinity with institutionalist perspectives. Furthermore, there might be an intimate relation between learning theories and evolutionary theories in the sense that learning is one mechanism through which diversity is created. Learning might even be an element in processes of selection.
Carlsson and Stankiewicz, Nelson and Rosenberg as well as Lundvall and his colleagues are all committed to the idea that technological change is an evolutionary process. Not only is the Systems of Innovation approach compatible with evolutionary theories of innovation but it could even be argued that there is a close affinity between the two.44 This is also true of interactive learning theories of innovation. In addition, both evolutionary and interactive learning theories are connected, either implicitly or explicitly, with institutionalist perspectives. These connections appear to be strongest with respect to their common emphasis on the importance of non-market mechanisms for reproduction, the generation of diversity, and selection (Edquist and Johnson 1997) . Hence theories of interactive learning together with evolutionary theories of technical change constitute origins of the Systems of Innovation (SI) approach, and both are certainly compatible with it.45 So, too, are institutionalist perspectives.
4.3. 'Systems' Perspectives on Interdependencies and Interaction
Having described the Systems of Innovation (SI) approach(es) in general terms, we will now proceed to a more detailed discussion of those specific theoretical contributions that are particularly important to the understanding of Government Technology Procurement (GTP) as an instrument of innovation policy. The later parts of our discussion will focus on current developments in SI or SI-related theorizing that are especially relevant to this topic. We will begin, however, with a survey of theoretical antecedents of what was later to be called the Systems of Innovation (SI) approach(es). We will concentrate on earlier theoretical perspectives on innovation in which
44 & " ' " , " + 5 ,' ' A " 8 , 5 / 9 5 / ' = 1 9 5 " " " * + 6 % " " 6 % 45 E " ' ' 8 " ' ; . ; . 5 69 interdependencies and interaction between users, producers and other actors were considered to be of central importance.
There are numerous antecedents of the Systems of Innovation (SI) approach(es). Those which are particularly relevant to the understanding of Government Technology Procurement (GTP) include, as we have suggested, a number of different but complementary theoretical perspectives on user-producer interaction. Some of these have already been mentioned. However, there is also a large body of theoretical and empirical work on the subject of what have been called 'innovation networks' (Freeman 1991) that is of relevance to our topic. In addition, there is an important tradition of theory and research on vertical integration and its relation to economic development (Perroux 1969a) that has strongly influenced the Systems of Innovation (SI) approach(es). One example of such influence is the use of the concept of 'development blocks' (Dahmén 1970) to understand the formation and development of the different systems of innovation in Denmark and Sweden (Edquist and Lundvall 1993) . We will deal with each of these literatures in turn, focusing primarily on the main implications for the analysis of Government Technology Procurement (GTP) within a Systems of Innovation (SI) perspective.
4.3.1 Chain-linked model
We have already referred, in our initial discussion in section 4, to the chain-linked model of innovation as an early systems-oriented perspective on innovation and precursor of the SI approach(es) (Kline and Rosenberg 1986) . This systemic view of the innovation process was counterposed to the conventional 'linear model' and some of its main features were described. In resuming this discussion, we should emphasize that the differences between the linear and chain-linked models of innovation are not 'simply' theoretical. The chain-linked model was developed primarily on the basis of extensive empirical evidence, much of it in the form of historical studies of technological development (e.g., Rosenberg 1976; Rosenberg 1982) and contemporary case studies of the management of innovation (e.g., Tushman and Moore 1982) , that contradicted various aspects of the linear model.
The chain-linked model was an important step in the development of what later became known as Systems of Innovation (SI) approach(es) for several reasons. As we have already mentioned, this was largely because it emphasized the importance of, and identified the possibilities for, multiple kinds of interaction among the various actors in the innovation process. These ideas were expressed in terms of alternative forms of 70 linkage among the several identifiable stages of development in a process of innovation, and pointed to the existence of several pathways important to innovation processes. These included not only the central 'chain' of innovation, but also numerous feedbacks, coordinating R&D with production and marketing, side-links to research all along the central chain, the importance of long-range generic research as a back-up for innovation, potentiation of wholly new devices or processes from research, and the infrastructural support (as well as stimulation and focussing) lent to science by technology (Kline and Rosenberg 1986: 303) .
There were, however, several other aspects of the chain-linked model that made it particularly relevant to the topic of government technology policy, and, hence, Government Technology Procurement (GTP). The most centrally important of these aspects were its strong emphasis on the crucial role played by the 'demand-side' in the innovation process and its closely related insistence that "Contrary to much common wisdom, the initiating step in most innovations is not research, but rather design" (Kline and Rosenberg 1986: 302) . A bias towards product markets and product innovation was suggested by these terms of reference. A policy perspective was also implicit in these features of the model, although it was not expressed as such. Rather, the authors spoke in terms of the "management of technology":
Larger uncertainty is strongly correlated with the degree of change. In the early stages of a product's life-cycle, major changes in product design are occurring rapidly, and the key problem of management is to find dominant successful designs and to organize stable production and marketing around them. ...
Some organizations are very effective in high-risk, radical innovations, others in the small, cumulative evolutionary changes that reduce costs and bring better fit of the product to various market niches. Both types of innovation are important. ...
In this connection, the very high costs for the development of new products, the shortening product life-cycle times, and the forces tending to squeeze out independent entrepreneurs in some heavy industrial sectors all suggest that the United States may need to rethink the way it has financed and managed innovation in some types of cases. (Kline and Rosenberg 1986: 303 - 304)
Implicit in these statements is the perspective that recognition of the complementary strengths of different types of firms and effective coordination among these and other actors will be essential to the creation and development of viable "innovation chains" for the design, production and marketing of new products, and that government may 71 have an important role to play, either directly or indirectly, in this capacity. This perspective is quite different from that of standard economic theory -- which, as the theoretical basis of the linear model, has tended to neglect both product innovation and the structural character of interfirm relationships (Andersen 1992: section 4.5.1) .
The primary focus of mainstream economic theory has been on process improvements achieved through "learning by doing" in competitive markets where there are no fundamental differences among firms (Arrow 1962b) . Relationships between buyers and sellers are here assumed to be standardized to the extent that price signals are sufficient to convey all of the relevant information required by the parties to any transaction (Hayek 1948) . In contrast to this orthodox view of an "extreme degree of flexibility" in the relationships of the economic system, the chain-linked model prefigured the Systems of Innovation approach(es) in devoting much greater attention to issues of "stability" and "'linkage' structure" (Andersen 1992: 82) .
4.3.2 Distributed process model
The notion of "chains" of innovation involving structurally differentiated roles for firms of different kinds (i.e., with different competences or capacities) was complemented by von Hippel's (1988) elaboration of an empirically grounded theory of variation in the functional sources of innovation. This theory challenged the long-standing assumption that "product innovations are typically developed by product manufacturers" and replaced it with a model of a "distributed innovation process" in which product innovations could originate from any one (or combinations) of at least three distinct (functional) sources: suppliers, producers and users (von Hippel 1988: 3 - 5) . The theory held that the "source" of innovation would vary according to the share of temporary benefits, or "economic rents", that firms in each category could expect from a potential innovation (von Hippel 1988: 6, ch. 4) .
This theoretical formulation successfully explained why the sources of innovation often differ by category of innovation. It explained, for example, why major product innovations in fields such as scientific instruments are almost always developed by product users (von Hippel 1988: ch. 2) , and why, in contrast, either product manufacturers or suppliers are the developers of most important innovations in other fields (von Hippel 1988: ch. 3) . The further development of this theory required detailed investigations of how firms formulate their expectations of profit from innovation. These investigations, in turn, indicated that the underlying principles often have to do with the nature of the relationship between the category of potential 72 innovator and the type of innovation being considered (in cases of process innovation, for example, users can profit from secrecy, while manufacturers and others can not) (von Hippel 1988: ch. 5) .
In exploring strategic implications of the effects of incentive structures on "distributed innovation processes", the theory pointed to the existence of non-rent-seeking practices of innovation such as "informal cooperative R&D" (or "know-how trading"). This occurs in industries where no firm among a group of rivals possesses all of the information required for innovation and none of them has sufficient incentive to develop a part of this information on an independent basis (von Hippel 1988: ch. 6, 90) . Finally, the theory identified viable strategies (such as the "lead user concept") for "shifting the sources of innovation". This term meant altering the normal patterns of interaction and information exchange among users and producers within a given product market so that the modified incentive structure resulting from these changes would induce more rapid and succesful product innovation (von Hippel 1988: ch. 8) .
Clearly, the theory and model of a "distributed innovation process" was an important forerunner of the Systems of Innovation (SI) approach(es). The strong emphasis on user-producer interaction and the interest in identifying and explaining the different structural configurations that could be assumed by such interaction implied a definite systems orientation. This was made explicit in statements to the effect that "both innovation managers and government policymakers will find it useful to understand innovation behavior at the systems level" (von Hippel 1988: 120) . The most important systems-level implications for policy-makers, in particular, lay in the discovery that the locus of product innovation does not always reside with product manufacturers, but is, instead, 'moveable' -- and can, moreover, be strategically shifted by altering incentive structures and re-ordering relationships between users and producers.
The main policy example given in this connection was that of the U.S. semiconductor industry, which lost most of its market share in the production of 'leading edge' computer memory chips to Japanese competitors during the 1980s. Among U.S. policy- makers, it was alleged, the common wisdom was that U.S. semiconductor process equipment firms "should somehow be strengthened and helped to innovate so that U.S. semiconductor process equipment users (makers of semiconductors) will not also fall behind" (von Hippel 1988: 9) . However, the finding that users, rather than producers, were the developers of most innovations in semiconductor process equipment (von Hippel 1988: Table 1-1) , indicated that the problem had a reverse causality and suggested that it could best be remedied by assisting U.S. users to resume innovation at the 'leading edge'. The following solutions were proposed: 73
We may, for example, think about lowering the innovation-related costs user-innovators expect. One approach to this would be for government funding for catch-up process R&D directed to advanced U.S. user firms and targeted to leading-edge process problems. Another would involve encouraging user firms to lower R&D costs by entering into cooperative work on process development. On the benefit side, we may think about how to make innovators' expected profits from in-house process development more certain. Could one, for example, make it advantageous for chip buyers to negotiate firm orders for major purposes of chips requiring new process technology far in advance? This would allow users a more certain return on their innovation investment. (von Hippel 1988: 122)
It can be seen from this example -- and more generally from the principles involved in analysis of the specific case under consideration -- that a systems-level understanding of "distributed innovation processes" can generate both supply and demand-side solutions to innovation problems, and that these different types of solutions need not be mutually exclusive. Of particular interest to the present topic, however, is the degree to which the demand-side solutions proposed in this case would necessitate a reorganization of the existing structure of demand. Moreover, it is worth emphasizing that the desired result of these measures would be to create (or re-create) a specific kind of relationship, conducive to further innovation, between the users and producers of semiconductor process equipment. That is, one where (as previously) "U.S. equipment manufacturing firms who, owing to geographic proximity to the innovating U.S. users, were typically the first equipment builders to gain access to the innovation" (von Hippel 1988: 121) . Thus, while Government Technology Procurement could have been one instrument for implementing the proposed demand-side (or "benefit side") solution in this case, its use would have implied a broader purpose of reorganizing the U.S. semiconductor and semiconductor process equipment industries at a system level.
4.3.3 Interactive learning theories
A third precursor to the Systems of Innovation (SI) approach(es), already mentioned in our general discussion in section 3.1 is the body of theory and research on "interactive learning" that is most strongly identified with the Aalborg school (Lundvall 1988) , although it has had other important contributors (Cohen and Levinthal 1990) . In both the "chain linked" and "distributed process" models of innovation, there was a strong emphasis on user-producer interaction in product innovation -- for the former, in terms of multiple feedback paths and differences in competency within innovation chains; for the latter, in terms of how the locus or source of innovation could be altered by shifting 74 incentive structures. While both of these theoretical formulations made reference to innovation as a learning process, however, neither of them elaborated on the 'learning' dimension of user-producer interaction in product innovation in a systematic way. This was the distinctive contribution of interactive learning theory to what later became known as the Systems of Innovation (SI) approach(es).
The core elements of an interactive learning theory of user-producer interaction in product innovation were first fully articulated in an Aalborg working paper (Lundvall 1985) . In this paper, direct opposition was taken to standard economic theory's perspective on innovation in order to relate innovation as a learning process to the dynamics of 'organized markets'. The main point of departure was a critique of the orthodox view of innovation as a process of "learning-by-doing" that (ideally) takes place within firms situated in perfectly competitive markets where firm behaviour is adjusted according to signals in the form of prices and quantities (Arrow 1962b) . In such a world, it was argued, cost-reducing process innovation (i.e., learning-by-doing) induced by price signals is possible and can even take place simultaneously within all producer units, but "it should be obvious that perfect competition does not induce product innovation" (Lundvall 1985: 17) . This is because in an anonymous market, where all communication takes the form of price signals, producers can not acquire information about user needs not already served by the market -- just as, conversely, users have no means of assessing new products, especially complex ones. In the 'real world', the argument continued, product innovation is ubiquitous, comprising the majority of important innovations and indicating the existence of (non-anonymous) user-producer relations (Pavitt 1984) . This necessitates an explanation of product innovation as a phenomenon associated with 'imperfect markets'. In this connection, moreover, "a user-producer perspective introduces the need for qualitative information about new use-values used as inputs and about the needs of users" (Lundvall 1985: 69) .
Up to this point, the argument developed was essentially the presentation of a strong case for vertical integration. If perfect competition militates against product innovation by restricting access to information about user needs, the logical solution will be for a producer to integrate with a user. The producer will then be able to acquire the necessary information about new technical opportunities and the user will become more competitive in relation to other users. This dynamic undermines the anonymous market and "gives rise to concentration both on the producer and the user side of the market" (Lundvall 1985: 20) . However, the argument then observed, the product innovation made possible by a partial integration of users and producers and the concomitant creation of a market characterized by 'small numbers' would result in unacceptably high levels of uncertainty and potential for 'opportunistic behaviour', since innovating units 75 would have much better information about new products than potential users. Trends towards vertical integration would be further reinforced, and complete vertical integration in the long run would reproduce one of the main effects of perfect competition: "All innovations should be in-house process innovations" (Lundvall 1985: 68) . For product innovation to occur, there must, therefore, be limits to vertical integration. These arise from the competitive need for both users and producers to maintain broad access to information about product capabilities and user needs, respectively.
The answer posed to this apparent paradox was the solution of "organized markets". This solution would involve a regulation, by mutually agreed to (i.e., consensual) "codes of conduct", of economic exchanges between users and producers. It would diminish opportunities for "cheating" and reward "trustworthiness" while increasing the possibilities for the flow of information between users and producers. The quality of such information, moreover, would be improved through coordinating mechanisms such as user associations. In the organized market, then, trustworthy producers would be rewarded by users with reliable information, and relatively stable relationships would develop between users and producers. This theoretical solution presented a world in which product innovation was made possible by rendering the market 'non-anonymous' while at the same time avoiding the pitfalls of complete vertical integration: "This vertical 'semi-integration' ... is a more easily reversed relationship that will not have as strong a negative impact on the flow of information as full integration" (Lundvall 1985: 28) . Thus, it was a solution which allowed for the existence of elements of hierarchy, but demanded the coexistence with these of countervailing elements of cooperation -- i.e., "mutual trust and responsibility" (Lundvall 1985: 29) .
The policy implications of this theoretical formulation only became apparent when the 'learning' dimension of organized markets was brought into clearer focus (Lundvall 1988) . In markets where products were simple and relatively inexpensive, the theory did not expect either organization or learning to be pronounced, but it anticipated that both would be important in markets where products were complex and their use-value characteristics were changing rapidly. In such cases, the requirements for direct cooperation and exchange of qualitative information between users and producers would lead to the establishment, through such interactions, of specialized channels and codes of information (Arrow 1974) . The strong association between complex technologies and differentiated communications systems had, in fact, already been well established in organizational research on innovation (Burns and Stalker 1961; Tushman 1977) . The relationship to learning, both within and among firms, was soon to be expressed in the concept of 'absorptive capacity', which referred to the patterning of 76 communication linkages between organizations and their environments, and its internalization or reproduction in the distribution of expertise within organizations, in terms of "cognitive structure" (Cohen and Levinthal 1990: 131 - 133). The authors of this concept were explicit that both the knowlege of firms and their capacities to acquire and exploit knowledge were the products of 'learning by interacting', and that their qualitative aspects were governed by the patterning of interaction:
Beyond diverse knowledge structures, the sort of knowledge that individuals should possess to enhance organizational absorptive capacity is also important. Critical knowledge does not simply include substantive, technical knowledge; it also includes awareness of where complementary expertise resides within and outside of the organization. ... To the extent that an organization develops a broad and active network of internal and external relationships, individuals' awareness of others' capabilities will be strengthened. As a result, individual absorptive capacities are leveraged all the more, and the organization's absorptive capacity is strengthened. (Cohen and Levinthal 1990: 133 - 134)
Being based on selective relationships involving elements of hierarchy and trust, and only capable of realizing their effectiveness through learning processes extending over relatively long periods of time, these acquired "knowledge structures" could well become sources of inertia and resistance to change. Although they were necessary for innovation, their relational character gave rise to an inherent conservativism: "The investment in information channels and codes will be lost if the old relationships are severed and new investment in the creation of new relationships will be required" (Lundvall 1988: 355) . Such problems would be especially pronounced when the emergence of a new technological paradigm (Dosi 1982) rendered established norms and standards obsolete and made old codes incapable of communicating information about innovative activities. And, more generally, "'effectiveness' does not ... guarantee efficiency if the criterion is user satisfaction at a low cost" (Lundvall 1988: 356) .46
While the Aalborg school viewed this problem of structural inertia primarily in terms of the self-reinforcing dynamics of established patterns of inter-organizational relationships, it could also be discussed in terms of the cumulative and path-dependent character of intra-organizational knowledge structures (Cohen and Levinthal 1990: 135 - 138) . If prior knowledge permits the assimilation of new knowledge, and some portion of this prior knowlege must be diverse, although still related, in order to permit the use of new knowledge for innovation, the development of 'absorptive capacity' over
46 0 8' " "" ! ! " "" ! 3 ! 77 time will tend to follow a definite trajectory requiring periodic readjustment and constant renewal. Without continuous development of this kind, the result may be "lock-out" from rapidly evolving fields (Cohen and Levinthal 1989) . "If firms do not invest in developing absorptive capacity in a particular area of expertise early on, it may not be in their interest to develop that capacity subsequently, even after major advances in the field" (Cohen and Levinthal 1990: 137) .
The other side of this coin has been referred to as "lock-in" -- a "competency trap" in which "favorable performance with an inferior procedure leads an organization to accumulate more experience with it, thus keeping experience with a superior procedure inadequate to make it rewarding to use" (Levitt and March 1988: 322) . "Lock-in" can be usefully analyzed in terms of the absence of internalized 'triggers' to innovation in the form of recognized failure to meet well defined aspiration levels:
If the firm engages in little innovative activity, and is therefore relatively insensitive to opportunities in the external environment, it will have a low level of aspiration with regard to the exploitation of new technology, which in turn implies that it will continue to devote little effort to innovation. This becomes a self-reinforcing cycle. Likewise, if an organization has a high aspiration level, influenced by externally generated technological opportunities, it will conduct more innovative activity and thereby increase its awareness of outside opportunities. Consequently, its aspiration level will remain high. (Cohen and Levinthal 1990: 138)
A further implication of this argument, of course, is that higher and more definite aspiration levels can be imposed from external sources -- namely, users. It was from this vantage that interactive learning theory considered the various kinds of 'unsatisfactory innovations' that had arisen from established forms of user-producer interaction in organized markets. These were most often cases of producer-dominated innovative activity deviating from user needs. Producer-dominated fields such as consumer goods, where both the information exchange with users and the innovation process were organized by producers provided a classic example (Freeman 1982: 202ff) .
Another example was the 'hyper-automation' of the Danish Dairy industry, where there was an assymetrical relationship between a few large producers of complex and technologically sophisticated equipment and a wide population of small, independent users (Lundvall, Maarbjerg-Olesen, and Aaen 1983) . Correction of cases such as these called for a pooling of competence by users, possibly reinforced by government regulation or support for user organizations. In some historical instances, though, users' commitment to existing user-producer networks, resulting in their resistance to the 78 formation of linkages with the producers of emerging technologies, could lead to much broader economic problems of 'mismatch' and reduced productivity that could not be resolved simply by strengthening conservative users (Freeman and Perez 1988) . These considerations led to the formulation of a general perspective on innovation policy, based on interactive learning theory's principle of "emphasis more upon the quality of demand, than upon demand as a quantitative variable" (Lundvall 1988: 357) .
The policy perspective of interactive learning theory reinforced and extended the policy implications of the "chain linked" and "distibuted process" models of innovation. It, too, suggested that the important role played by demand in provoking innovations should not be taken as a justification for laissez-faire policies but, rather, as the basis for strategic intervention in the demand side. Its first main consideration was that "The lack of competence by users and the tendency of producers to dominate the process of innovation might be as serious a problem as the lack of competence on the producer side" (Lundvall 1988: 358) .
The second main consideration was that "government may intervene, directly or indirectly, in relation to the establishment and restructuring of user-producer relationships" (Lundvall 1988: 358) . In periods of gradual change, for example, government could either sustain existing linkages or reorder them by creating new organizations that would intermediate between users and producers, condense and improve the distribution of information, and thereby stimulate 'satisfactory' innovations enhancing economic welfare.
The third main consideration was that in periods of radical technological change, government will need to transform many established user-producer relationships in order to overcome inertia in organized markets supported by vested interests in the existing structure. "The difficult task for government will be to stimulate the renewal, or severance, of well established user-producer relationships and the establishment of new relationships" (Lundvall 1988: 358) .
All of these three considerations pointed very directly to Government Technology Procurement as a policy instrument enabling government to act either directly as a user or indirectly on behalf of users in the development of innovations and the creation of 'organized markets' conducive to innovation.
4.3.4 Network analysis 79
As we have seen, the strong emphasis on institutional analysis in interactive learning theory led to the identification of a broad realm of economic relationships and innovative activities that belonged to the domain of "neither market nor hierarchy". This, in turn, paved the way for another important influence on Systems of Innovation (SI) approach(es) -- namely, the growing body of theory and research on innovation networks, or 'networks of innovators' (Freeman 1991) . There is now a voluminous literature on networks as a special form of economic organization between markets and hierarchies (see, for example, Johansson 1991; Perrow 1993; Powell 1990) , and much of this work has dealt in various ways with themes such as 'learning' and 'innovation' (see, for example, De Bresson and Amesse 1991; Imai and Baba 1989; Matthews 1996) . It is beyond the scope of this paper to provide a comprehensive review of this literature. Instead, we will focus selectively on contributions that may be considered as particularly important formative influences on the Systems of Innovation (SI) approach(es), and on those which are of special relevance to the understanding of Government Technology Procurement (GTP).
Håkansson's (1989) work on industrial networks has been cited as an important influence on at least two major variants of the Systems of Innovation (SI) approach (Carlsson 1995; Gelsing 1992, in Lundvall 1992) . The affinity between network and systems analyses is evident in Håkansson's summing up of the network approach. According to him, network analysis: "focuses on events and developments between companies and organisations, rather than ... looking at events chiefly within the companies"; it stresses "specific relationships ... as opposed to ... general relations between international, national and regional developments"; and, it is concerned with dynamics, since "A network is never stable or in balance, but is always changing in all kinds of ways" (Håkansson 1989: 170) .
Despite its very strong empirical orientation, this particular body of research and analysis has made important contributions to the further development of 'interactive learning', 'chain-linked', and 'distributed process' models of innovation and it has done so in a manner relevant for the understanding of Government Technology Procurement (GTP) as a policy instrument. Other, more theoretically oriented contributions to the literature on innovation networks (in particular, Teubal, Yinnon, and Zuscovitch 1991) , that have made this relevance more explicit, will be discussed below. For the moment, however, we will dwell upon some empirical generalizations drawn from Håkansson's (1989) study of technological collaboration in industrial networks as later discussed by Håkansson (1990) . 80
As we have seen, there is a tension between vertical and horizontal interdependencies, or forms of integration, implicit in all of the systems-oriented models of innovation considered thus far. This tension becomes explicit in interactive learning theory's treatment of the opposition between 'markets' and 'hierarchy'.47 Not surprisingly, therefore, a major theme in the exploration of the organizational architecture of industrial networks devoted to collaborative development of new technologies, has been the investigation of 'vertical' and 'horizontal' relations within networks. The empirical work under discussion here (a study of 123 innovative Small and Medium-sized Enterprises, or SMEs, in middle Sweden) discovered a dynamic relationship between these two dimensions of network structure (Håkansson 1990) .
In this description and analysis of networks, 'vertical ' and 'horizontal' forms of inter- firm relationships were conceptualized somewhat differently from, but in a manner consistent with, the Aalborg version of the 'interactive learning' perspective discussed in subsection 3.3.3. This is because all user-producer (or customer-supplier) relations constituting inter-firm networks must, by definition, involve some degree of vertical integration (although, again by definition, vertical integration can never be 'complete' in the case of an inter-firm network). Networks of this kind are, in other words, 'organized markets' -- they are fundamentally different in this respect from the 'anonymous markets' that are used by the Aalborg school to represent a pure form of horiziontal integration. It follows that in network analysis the use of the term 'horizontal' must be reserved for discussing non-anonymous and collaborative relationships among firms other than those which are based on the sale and purchase of goods or services (user- producer relations). In network analysis, therefore, the term 'horizontal' is used to describe networks based on other kinds of relations -- including, for example, the "cooperation between rivals" or "informal know-how trading" discussed by von Hippel (1988). (See subsection 4.3.2). This usage generated a basic distinction between 'trade' networks with a strong vertical aspect and 'knowledge' networks with a strong horizontal aspect (Gelsing 1992: 120) .48
47 ' 8 * " " ! 3 !' ' ! ! * " " ! !' * " 3 48 " " 8 8 8 8 " " " " " " 8 8 " " " " " " " 8 8 " * 6 "99$$*7 81
Håkansson's (1989) research found a significant correlation between 'proportion of customer-led production' and 'external share in technical development' (Håkansson 1990: Table 4) . 49 This demonstrated that "Customer-led production gives the company a general outward-orientation which is echoed in its technical development activities" (Håkansson 1990: 375) . On one hand, this finding showed the central importance of 'vertical integration' in the form of 'user-producer' relationships for product innovation within individual firms.50 On the other hand, the association of 'customer-led production' with an 'outward orientation' in technical development indicated a dynamic towards market diversification or market-creation on the part of innovative firms and groups of firms. Partial vertical integration, essential for product innovation, did not preclude but rather favoured a search for new markets.
This virtuous relationship was also demonstrated in another way, with respect to 'horizontal' relationships. So-called "horizontal cooperation partners" -- a diverse group including complementary producers, rivals, universities and other research institutions - - were also found to be of considerable importance to firms (Håkansson 1990: 376 - 377) . Although "horizontal" development relationships were statistically less frequent and were typically of shorter duration, they were more directly linked to a conscious selection of relationships and a primary interest in specific collaboration projects intended not only to enhance the existing resources of firms but also to "cultivate ... their environments" (Håkansson 1990: 378) . One implication was that these kinds of relations, formed primarily for the purpose of product innovation, were intended to
49 ! ! " ' ' " E" "' B C# 77 77 C E" ' B C # ' 7 7 3 " - C @ C+ ' " 77 ' ' ' " * C @I8 =B %5$ ' ! ! " " " @' " " " " * " 8 !7 !' " " " 8 - " "' * ' " " " ! ! " "" 7 % % % 50 @I8 =B : " " C 8 5:D " C' C C @I8 =B %5 82 develop both new markets and new 'vertical' development relationships (user-producer relations).51
These two aspects of network evolution through 'horizontal' technological collaboration -- the development of 'absorptive capacity', or learning, and the search for new 'vertical' user-producer relationships, or market creation -- were the focus of Teubal, Yinnon and Zuscovitch's (1991) work on the theoretical and policy implications of network analysis. As these authors pointed out, the primary difference between their theoretical approach and that of Lundvall's (1983; 1988) work on user-producer relationships could be defined primarily in terms of policy relevance.52 Their central theme was the relationship between networks and market creation, with particular reference to "the user-producer relationships characterizing ... many emerging capital-goods markets" (Teubal, Yinnon, and Zuscovitch 1991: 381) . Their thesis was that the emergence of a vertical form of network centred on a focal organization could be crucially important in "linking a radical capital-good invention to its subsequent contribution to the national economy" (Teubal, Yinnon, and Zuscovitch 1991: 382) .53
These authors argued that one of the main distinctions between novel capital goods networks and other types of network formations (such as supplier-assembler networks in automobile manufacturing) is that the former, do not, like the latter, tend to possess strong central agencies (such as the focal auto-assembly firm). Therefore, they often
51 " 3 C C " " " 8B
0 3 ' " " " " " " J 0 7"" " ' " " " " ' " " " "
1 3 ' / " "" "" " ' ' " @I8 9 %
52 ' " C C ! 3 8 !' C " 8 ' C ' K' L B %A$' "" 53 / " 3 ' ! ! " 8 ' 8 " 3 " " 8 " " " C M C ' K' L B %A$ 83 lack the means of centralized communication between producers and a broad range of users (Teubal, Yinnon, and Zuscovitch 1991: 382) .54 The absence of an agency capable of these functions not only slows the pace of product innovation dependent upon user-producer interaction -- i.e., learning -- but it also limits producers' capability to match product types to new users -- i.e., market creation.55
The argument then proceeded to consider whether market relations would "naturally" permit the emergence of a central agency that would carry out the tasks of organizing 'collective learning' by linking users and 'articulating demand'. The problem facing an entrepreneurial supplier who might assume this role, it was argued, was that it would involve "direct coupling with the needs of an expanding set of users over and beyond the initial application" (Teubal, Yinnon, and Zuscovitch 1991: 386) . To overcome this problem, the entrepreneur would require extraordinary capabilities with respect to the coordination and scope of R&D in a 'new market' environment where opportunities for the expansion of sales through the gathering and diffusion of relevant information would be extremely difficult.56
This "critical mass" problem, it was argued, was only likely to be resolved by the early formation of "a core of innovative users ... participating together ... in the innovation process as a whole" (Teubal, Yinnon, and Zuscovitch 1991: 389) .57 While markets might eventually permit this to occur at a later point, they were not very likely to allow its occurrence 'early in the game'.
On this basis, a case could be made for government intervention in the process of market creation -- especially where the social profitability of a given technology was
54 ' " " " " " C " 8 * " ' "" " 7 " * C ' K' L B %A$ 55 " " " 8 ! 3 ! * " 8 ! ! " " " ! ! "" 8 !" ! ' C 8 " C ' K' L B %A< 56 "" " ' !! 8 ' C "' 8 7 ! ! 1 " 8 C ' K' L B %A5 57 + " " + * ' ' ! ! " " "" ! ! * 84 clear from comparative evidence. Essentially, what was required to resolve the coordination problems involved in market creation was an articulation (or organization) of demand resulting in, or equivalent to, the formation of a sufficiently large core of innovative users.
Government, it was reasoned, might accomplish this indirectly, through technology diffusion and market development programs (Teubal, Yinnon, and Zuscovitch 1991: 390) .58 It was a relatively short step to the consideration of a more direct role for government 'articulation of demand' in the cooperative development of new generic technological capabilities requiring the participation of competing users. Cases of this nature, it was argued, "suggest a shift in the role of government", especially where cooperation is difficult and users are competitors, and there is no tradition of cooperative development: "Thus government's role may be in some cases that of ... orchestrating and even initiating and partially financing such attempts" (Teubal, Yinnon, and Zuscovitch 1991: 391) . The type of policy role being advocated for government agencies here was, in other words, broadly similar to the 'network coordination' role played by MITI in the 'Japanese model' of GTP discussed in section 1.1.
4.3.5 Development block theory
As we have seen, theoretical analysis of innovation networks has posed problems regarding the coordination of horizontal patterns of interaction in technology development. The resolution of these problems depends on focal agencies or mechanisms around which new 'organized markets' -- in Lundvall's phrase, forms of 'vertical semi-integration' -- can be built. These observations lead, naturally, to a further consideration of the theme of 'vertical integration' and its importance within the Systems of Innovation (SI) approach(es). As stated earlier, this analytic focus has been carried into the Systems of Innovation (SI) literature primarily in the form of theoretical and empirical work on "development blocks" (Andersen 1992; Carlsson and Stankiewicz 1995; Edquist and Lundvall 1993) -- a concept originated by the economic historian
58 * * " + ' , 3 C" "" " " " 8 " " "C ' K' L B %= + ' C ' ' " "' " C C " ' " " " " ' " 8 C ' K' L B %= 85
Erik Dahmén (1970) , and closely associated with the Perrouxian concept of 'growth poles' (Perroux 1969a) . Development blocks are similar to what Michael Porter (1990) would later call 'clusters' and what others have labelled industrial complexes.
Growth pole theory originally dealt with the impact of radical innovations on economic structure (Perroux 1969a) . It advanced a thesis concerning waves of structural economic change, initiated by industries that were primary or autonomous innovators, and carried through by industries that were adaptive or induced innovators (i.e., impelled by the first group). However, later applications of the theory in regional and development economics tended to disregard both its original conception of economic space (Perroux 1969b) and the fundamental point that the essential basis for the creation of a growth pole was not simply the presence of a certain type of industry, but rather that its activity was radically innovative (Brookfield 1975: 93) . For these reasons, the theory lost credibility and eventually fell into disuse (Andersen 1992: 70 - 71) .
The theory of development blocks has fared somewhat better, having been incorporated, more or less intact, into Systems of Innovation (SI) theorizing. One important reason for this may be that the theory was resonant with many of the precursors of the Systems of Innovation (SI) approach(es) that have been reviewed here. It was well articulated, in particular, with network theories and models of user-producer interaction, to which it added a special emphasis on the role of 'structural tensions' in economic development. These aspects of the theory were evident in the way a "development block" was defined. Abstractly, it referred to changing relationships among economic agents as "a sequence of complementarities which by way of a series of structural tensions, i.e., disequilibria, may result in a balanced situation" (Dahmén 1988) . More concretely, the concept has also been discussed in terms of a broad set of interconnected producers and users of products, developing in close interaction with each other and often supported by knowledge-producing private, or public, organizations -- including in its most complete form whole chains of production, final users of the products, and producers of the means of production, as well as independent knowledge resources (Stenberg 1987: 46 - 47) .
Parallels with the concepts of vertically integrated "chains of innovation" and "organized markets" or "networks" based on "user-producer interaction" are obvious, and require little further comment here. What is of particular interest to the present discussion is the structural and developmental perspective that the theory of development blocks brought to bear on the subject of 'entrepreneurship' in the context of technological innovation. As we have seen from the preceding discussion of the 86 evolution of networks formed in relation to innovative products, there is a basic developmental problem of 'critical mass' to be overcome, the solution of which requires vertical integration and coordination of information flows between users and producers through the centralizing agency of an exceptionally capable 'entrepreneur '. What 'development block' theory adds to this is the perspective that this need for linkage and coordination between users and producers arises in the context of structural economic tensions emanating from 'gaps' in technological development:
Uncompleted blocks are usually indicated by price and cost 'signals' on various markets. This could mean low current profits, or even losses in some areas but also promising prospects, if the steps necessary to complete the blocks are within reach. Less anonymous impulses may also come from actors in economic life, where, as we know, there are widespread networks of relations and contacts outside what theoreticians call a 'market'. In both cases, the challenge is in 'gap filling' which tends to eliminate structural tensions but may also lead to new tensions by overshooting, as technical and other solutions sometimes run ahead of the immediate goal. (Dahmén 1988: 6)
According to this formulation, development blocks can be discerned not only ex post, but also ex ante. On one hand, this feature permits the intervention of an entrepreneur; on the other, it makes possible the identification of the need for entrepreneurial intervention, even where a 'natural' entrepreneur does not exist. Potential development blocks can be identified because such a block "will, typically, involve several domestic sectors, coupled by strong quantitative and qualitative linkages" that form "frameworks for interactive learning" (Edquist and Lundvall 1993: 276) . The core of such a block will typically be a radical technological innovation that raises dramatically the degree of disequilibrium within the economic system, starting a sequence of structural tensions and partial resolutions -- that is, a progression of sets of new market niches that will continue until the development potential of the core technology is finally exhausted (Dahmén 1988) . Several different contexts in which emergent development blocks might be discerned have been discussed by Andersen (1992), who concludes that this type of analysis can be a useful guide to the formation of economic policies at the national level:
Here we ask whether the overall 'business situation' is characterised by a whole set of immature but strong development blocks (where new innovative opportunities continue to pop up) or whether the dominant development blocks have reached a set of saturation and, perhaps, overaccumulation. The first situation implies a positive 'transformation pressure' on the economy while the second sooner or later implies forced adaptation or creative destruction. 87
With such a description of the 'business situation' with respect to innovation, adaptation, and destruction it becomes possible to evaluate macroeconomic policies from a Schumpeterian stand-point. (Andersen 1992: 89 - 90)
As suggested by the earlier discussion of networks formed in relation to new technology, the 'natural' entrepreneur whose intervention is necessary to lift the network out of a 'low-level equilibrium trap' may often fail to arise, and this may necessitate government intervention. Essentially, government is then required to act 'in place of' such an ideal entrepreneur, to perform the complex role of articulating demand, eliciting appropriate producer response, and creating sufficient 'critical mass' to spur further development. 'Demand-side' intervention of this kind can be readily justified in cases where it is clear that further development of the technology in question would benefit economic welfare. Operationalizing the concept of 'development blocks', we have suggested, makes possible the identification of such cases. It directs the attention of economic policy makers to instances where there may be a 'transformation' problem between user needs and production characteristics in the early stages of development of a new product or industry. Further, it points to the need for entrepreneurial activity directed towards resolving this structural tension in a manner that will both stabilize the situation and open up new possibilities for development.
Government Technology Procurement (or GTP) is of course, one important means through which this might be accomplished where no 'natural' entrepreneur is present. This point has been clearly recognized in Systems of Innovation (SI) theorizing about the need for entrepreneurial competence in the transition from 'network' to 'development block':
Thus, the role of the entrepreneur is to provide the spark or the vision that turns a network into a development bloc. ... He (sic) has to perceive the (future) need, identify the necessary ingredients, secure the resources that may be missing initially, and communicate his vision to the relevant agents - - capitalists, suppliers of raw materials, people with the required skills, etc. ... In some cases, e.g., when the technology is complex and when there are many missing elements in the required resource set, the entrepreneur or "champion" needs to command large influence and/or financial resources, such as the U.S. Department of Defense, Japan's MITI, or a major business firm (first echelon firm or "lead user"). (Carlsson and Stankiewicz 1995: 40 - 41)
4.4 Summary: General Policy Implications of the SI Approach 88
In the foregoing sections (4 - 4.3) we have reviewed various theoretical aspects of the Systems of Innovation (SI) approach(es), concentrating on their relevance for the understanding of Government Technology Procurement (GTP) as an innovation policy instrument. Also, and particularly in the later part of the discussion (section 4.3), we have tried to outline some general policy implications of SI perspectives and their antecedents regarding Government Technology Procurement (GTP).
First, in section 4, we introduced a distinction between linear - and systems -oriented conceptions of the innovation process, demonstrating their affinities, respectively, with supply and demand-oriented innovation policies. Following our previous discussion (in part 2) of GTP as a policy instrument that is very clearly associated with demand- side innovation policies, we argued that GTP is not adequately comprehended by mainstream economic theory's 'market failure' explanations of the need for supply- oriented public intervention in the innovation process, nor is it adequately represented by the associated linear model of innovation. We suggested that GTP is better captured by systems oriented approaches to the understanding of innovation processes.
Second, (in sections 4.1 and 4.2) we provided an overview of the Systems of Innovation (SI) approach(es). We began (in section 4.1) by reviewing common characteristics of the various SI approach(es). We pointed out in this connection that non-linearity and interdependence are among the key characteristics. This makes it natural to include a focus on demand-side instruments in SI-based analyses of innovation policy. We went on (in section 3.2) to consider in more detail some important 'roots' of the Systems of Innovation (SI) approach(es) -- in particular, the influence by different theories of innovation such as interactive learning theories, evolutionary theories and institutional perspectives (Edquist 1997a: 8) . The discussion addressed aspects of SI approach(es) particularly relevant to Government Technology Procurement (GTP) as a matter of of interaction between users and producers (see also section 2). We argued in this connection that such interaction is not well understood by standard economic theory because it is not only a matter of price signals and quantities sold and bought. The interaction contains other kinds of information and also knowledge, and it can initiate learning processes. GTP is therefore better captured and more well understood by systems-oriented views of innovation.
Third, (in section 4.3), we discussed contributions to the Systems of Innovation (SI) approach(es) that focus upon interdependencies between elements (organizations and institutions) which influence processes of innovation. Our discussion dealt primarily with a number of important theoretical precursors of the Systems of Innovation (SI) 89 approach, although it also made reference to some of the SI literature. The SI 'antecedents' that were discussed were the 'chain-linked' and 'distributed process' models of the innovation process, 'interactive learning' theories, 'network' analysis, and the theory of 'development blocks'. All of these elements of what was later to develop into the Systems of Innovation (SI) approach(es) were shown to be relevant to the understanding of Government Technology Procurement (GTP) because of their articulation of different but complementary theoretical perspectives on interdependency and user-producer interaction.
Rather than review in any detail the theoretical aspects of the various complementary 'systems' perspectives on interdependency and user-producer interaction that were discussed in the previous section (3.3) we will now briefly summarize their main policy implications. Each of the antecedents of the Systems of Innovation (SI) approach(es) identified here can be associated with particular policy implications. However, these policy implications can also be taken as complementary components of a broader Systems of Innovation (SI) policy perspective -- in much the same way as the various 'appreciative' theories and models with which they were originally associated have since been incorporated into a broader set of theoretical perspectives on Systems of Innovation. What follows then, is a brief outline of those components of a Systems of Innovation policy perspective that are particularly relevant to the understanding of Government Technology Procurement (GTP):
1) The Chain-Linked Model -- placed a strong emphasis on the crucial role played by the 'demand-side' in the innovation process and developed a main focus on product markets and product innovation. It advanced the perspective that the 'management of innovation' (including public policy) should recognize the complementary strengths of different types of firms and seek to coordinate their efforts through the creation of viable 'chains of innovation' involving linkage structures among firms and other actors.
2) The Distributed Process Model -- identified different potential sources of innovation and revealed that the locus of product innovation certainly does not always reside with product manufacturers. Instead, it is 'moveable' and can be strategically shifted by altering incentive structures and re-ordering relationships between users and producers. This can be done, for example, by adopting a 'lead user' strategy to restore or advance the innovative competitiveness of producers. 'Lead user' strategies have a direct parallel in Government Technology Procurement. It should be recognized in this connection that both depend on the modification of existing incentive structures, and both imply a wider restructuring of existing relationships between users and producers. 90
3) Interactive Learning Theories -- developed a more general theoretical outlook on user-producer relationships in the context of product innovation, advancing a policy perspective that was based on the 'learning' dimension of user-producer interaction in 'organized markets'. This perspective was demand-oriented, but placed greater emphasis on the quality of demand than upon its simple quantity. It suggested that the important role played by demand should be taken as a basis for strategic intervention in the innovation process. Its first main consideration was that the tendency of users to dominate the innovation process might be as great a problem as any lack of user competence. Its second consideration was that government could intervene, directly or indirectly, to establish or restructure user-producer relationships -- for example by sustaining existing linkages or by supporting organizations formed to promote new linkages. The third main consideration was that in periods of sudden technological transformation governments may have to transform many established user-producer relationships to overcome inertia in organized markets reinforced by vested interests in the existing structure.
These aspects of 'interactive learning theory' pointed very directly to Government Technology Procurement as a policy instrument enabling government to act either directly as a user or indirectly on behalf of users in the development of innovations and the creation of 'organized markets' conducive to innovation This is not only relevant for "traditional" GTP but also for the 'NUTEK kind' of GTP, where the public agency does not represent final demand but serves instead as a catalyser.
4) Network Analysis -- developed further policy implications of the concept of 'organized markets' by pointing to the dynamic tension between vertical integration and horizontal technological collaboration among 'learning' organizations (users, producers, and others) in the creation of new product markets. A focal organization providing vertical linkage between users and producers was considered to be necessary to overcome a 'low-level equilibrium trap' facing networks formed around emergent technologies.
Markets, some analysts argued, were only likely to allow the emergence of a a focal 'entrepreneur' at a relatively late point in the development of a network based on new technology, given the many ex ante coordination problems that would have to be solved. Therefore, especially where the further development of a given technology could be shown to have beneficial consequences for economic welfare, a case could be made for government intervention to resolve the coordination problems involved in market creation. This would involve an organization or articulation of demand resulting in, or equivalent to, the formation of a core of innovative users sufficiently large to 91 overcome the problem of otherwise inadequate 'critical mass'. As pointed out above (under points 2 and 3), government could perform this role through GTP either directly, as a 'lead user', or indirectly, as a 'catalyser'. In both cases, government, as the 'focal' entrepreneur, would have to represent both the broad range of diversity in user demand and a high level of sophistication in 'user competence'.
5) Development Block Theory -- refers to innovation-induced 'structural tensions' giving rise, through disequilibria, to the formation of a broad set of interconnected producers and users of products, developing interactively, often with the aid of knowledge-producing organizations. A development block will include in its most complete form a whole chain of production, as well as independent sources of knowledge. It is a large-scale framework for 'interactive learning'. Essentially, these are vertically integrated (but still horizontally evolving) user-producer networks formed around radical technological innovations. They differ from 'networks' per se primarily in that they have successfully overcome the problem of 'critical mass'. Significantly, 'development blocks' can be discerned ex ante as well as ex post. Their identification and analysis can therefore be used to guide public policies for economic development based on technological innovation.
'Development Block' analysis enables economic policy makers to identify and evaluate 'transformation' problems between user needs and production characteristics occurring in the early development of a new technology. The underlying theory takes account of the need for 'entrepreneurship' to fill this 'gap' in a way that will both stabilize the situation and open up new possibilities for development. Government Technology Procurement (or GTP) is of course, one important means through which this might be accomplished where no 'natural' entrepreneur is present -- especially where the technology in question is extremely complex and the demand for resources and influence is large.
5. The Stages in a Process of Government Technology Procurement
Because the development of new products and systems is often very time-consuming, Government Technology Procurement is normally a long-term process. Often the process continues over several election periods, making it vulnerable to political decisions regarding continuation or termination of the project. The long maturation process also leads to the risk that the boundary between the purchaser and the supplier becomes obscured, a situation which is problematic for the management and control of the procurement process. It might also lead to problems of corruption. 92
These factors make it useful, for analytical purposes, to divide the process into phases. In the 'real world', of course, there may be many cases of Government Technology Procurement which do not proceed through the identifiable 'phases' in the same sequence. We might, for example, encounter some processes that conform to the following sequence:
Goal --> Tender --> Invention --> Development --> Testing --> Application --> Use -- > Diffusion
However, some other processes of GTP might instead proceed as follows:
Invention --> Development --> Goal --> Tender --> Diffusion --> Development --> Testing --> Application
It may be possible to identify other patterns as well. Perhaps the process started with the creation of research projects involving public agencies (potential procurers) and manufacturers (potential suppliers). Perhaps, the diviion of labour between the buyer and the manufacturer was not clear from the beginning (as in some 'ideal type'), but the result of an interactive process? As these examples suggest, there may be no 'ideal' or 'linear' scheme according to which procurement projects develop.
Notwithstanding these considerations, the identification of discrete phases may still be a useful device for describing and analyzing processes of Government Technology Procurement. For example, in the alternative sequences identified above, the same phases or stages can be recognized, though their order of progression differs. This provides a basis for comparison.
A technology procurement process can, according to Linder (1982:7) be divided into four phases: - a study phase, - a project phase, - a design phase. and - a production phase.
During the study phase the need which shall be satisfied or the problem which shall be resolved by the procurement is identified. The objective is made clear through identifying the function that the system or product shall perform. During the project phase this function is specified. During the design phase the solution is worked out in 93 detail and a prototype is often produced. After this, serial production follows during the production phase (Linder 1982: 7) .
Most procurement processes probably go through these phases in one way or the other, although certain overlaps can, of course, occur. This does not, however, imply that the actors involved are always the same in a specific stage. For example the problem/need identification in the study phase can be done by the potential supplier as well as by the future user (or by other actors, like politicians). The whole process might also start with the initiation of research projects involving procurers and also potential suppliers - as well as researchers.
The same is true for the project phase, where suppliers as well as users or researchers might be involved. In the design phase the importance of the future supplier has increased although the users may still have an important role. In the production phase, however, the supplier has taken over almost completely.
This means that there is no ideal scheme according to which all procurement projects develop. For example, the process might start with an identification of a problem and a functional specification of a system that could solve that problem. This could then lead to inventions and development work that makes the fulfilment of the specifications possible. This may then leads to design of the system and the production of it. Alternatively the whole process might start with an invention which makes someone realise that a previously unsatisfiable need can now be cared for. This need is then specified and the design work is initiated as a basis for future production. In other words, the division of labour between the buyer and the manufacturer is not necessarily clear from the beginning; the whole process might be initiated as a part of an interactive learning process.
However, at the same time there seems to be some degree of ’order’ in the procurement process, e.g. that the role of the supplier increases during the process. Government technology procurement is interactive but not anarchic. The role of the buyer is also much larger than what is indicated by the linear model of innovation. If anything, government technology procurement puts the linear model upside down, since the role of demand is so important.
6. The EC Procurement Rules 94
Although the EC procurement rules are not a matter of theory, they are a framework condition for most case studies. It is therefore convenient if they are adressed in this 'general' paper. However, national deviations from these rules must be dealt with in the various case study reports.
In this paper, an extensive description of the EC procurement rules has been provided in Appendix A. (See Appendix A.) The EC procurement rules have also been dealt with from the analytic perspective of European integration in section 1.3 and from that of 'auction theory' in section 3.3. Here, we will provide only a brief overview of the EC procurement rules, based on the account given in Appendix A.
The current EC procurement rules were developed in the early 1990s, as part of a larger effort to consolidate a single 'internal market' for public purchasing in Europe. The single market initiative grew out of frustration with lack of compliance with existing regulations and was primarily motivated by the hope realizing large savings in public expenditure, However, it was also argued in this connection that the opening up of hitherto protected public markets in areas such as railways, power generation, and telecommunications equipment would reduce over-capacity, and facilitate rationalisation in the form of mergers and investments in new technology. This would have "an important impact on the rate of innovation. investment and growth in the sectors enjoying the benefits of restructuring, with positive effects on their international competitiveness" (WS Atkins Management Consultants and Associates 1988: 7) .
Despite the professed concern with innovation, however, the EC procurement rules appear to have been designed "mainly for the purchasing of already developed goods or services, often standardized, and not specifically for development projects" (Westling 1996: Appendix 5, 1) . They are concerned primarily with procedures for source selection and the award of contracts, neither of which are dealt with exhaustively. Rather the rules provide a broad framework within which domestic laws can be applied. As of the nmid 1990s, utilities are covered by the rules, and coverage has been extended to services as well as supplies and works.
What is perhaps most significant for our purposes, though, is what the rules do not cover. This includes international projects, procurement governed by international agreements or involving the military, projects undertaken by international organizations. Justifiable exceptions can also be made for development projects "of a genuinely innovative nature for which the use of existing European standards, European technical approvals or common technical specifications would not be appropriate" [ c.f. \Westling, 1996 #295: Appendix 5, 5. 95
The rules prescribe three basic procedures of source selection: the open, restricted and negotiated procedures. The open procedure is one of unrestricted tendering, according to established rules of advertisement, closed bidding and contracting. It applies primarily to the purchase of existing, standardized supplies, works and services. Contracts can be awarded on the basis of 'the most economically advantageous offer' -- a criterion that takes into account not only cost but also technical merit, the reliability of the tenderer, and what might be termed transaction costs.
Development projects are more likely to be dealt with under the restricted procedure. This allows for invited tendering and more latitude for procurers (particularly utilities) in processes of pre-qualifying prospective suppliers and advertising contracts. It is justified by "a need to maintain a balance between contract value and procedural costs, and by the specific nature of the goods to be procured" (Business International Ltd. 1991: 27) .
Finally, there is the negotiated procedure, which is highly exclusive and is most likely to be pursued in highly innovative development projects. Innovation is, in fact, one of its primary justifications. The negotiated procedure can also be used for technical or artistic reasons, or to protect exclusive rights, when the goods can only be provided by a particular supplier. Additionally, its use can be justified in cases of extreme urgency, or where a change of suppliers would not be feasible for technical reasons. (Business International Ltd. 1991: 28)
The EC procurement rules are intended to eliminate artiificial barriers to trade, and this involves doing away with unnecessary differences in engineering standards and design specifications, particularly in high-technology supplying sectors. The existence of differing national design standards in areas like railways and power generation, it has been claimed in support of the rules, "creates a barrier to to the trade in goods and services supplied by the public sector, with widespread repercussions on transport and telecommunications, and hence on the efficiency of European business" (WS Atkins Management Consultants and Associates 1988: 4) .
The antidote has been to prevent buyers from using technical specifications to discriminate in favour of traditional national suppliers. To this end, the EC initiated in 1985 a 'new approach' to harmonizing technical standards that relies on a number of standard setting bodies to implement generally defined European Standards. Pending complete harmonization of technical standards, CEN, Cenelec and ETSI award products whose manufacturers have demonstrated compliance with essential requirements the 96
'CE' mark, which permits them (the products) to be sold freely throughout the EC (Business International Ltd. 1991: 54 - 55) . There are however, certain types of procurement which fall outside of the rules governing standards including both international projects and development projects of a genuinely innovative nature.
Until 1992 the EC's main mechanism for ensuring compliance with the procurement rules was the administration of remedies imposed by the European Court of Justice. This approach proved to be unwieldy and ineffective. In the 1990s, a more decentralized aopproach has been adopted, in which Member States administer Community Law. The rationale was to avoid "apoplexy at the centre and paralysis at the extremities" by achieving closer cooperation of Member States through the harmonization of national procedures for examining complaints (European Commission 1992) .
Under this new approach, national laws are to be harmonised with the framework provided by open-ended Directives (rather than highly prescriptive Regulations). The Commission, together with the European Court of Justice, remains ultimately responsible for monitoring and enforcement of compliance, but the bulk of this work is now carried out by national tribunals and regulatory authorities. Generally, the regime of enforcement relies upon unhappy enterprises to invoke remedies by submitting a complaint to a national tribunal.
The main conditions for effective remedies in the field of public procurement, as recognized in the EC procurement rules themselves, are: (i) fast review procedures allowing for early resolution of disputes, (ii) readily available interim measures permitting early resolution of disputes, and (iii) ensurance that adequate compensation will be paid to injured bidders in the event that early resolution is not achieved. The main purpose of of the directives concerning compliance and remedies was to correct insuffiencies with respect to all of these three conditions in existing national systems of redress.
However, that member states continue to enjoy considerable latitude with respect to all of the above-mentioned conditions. Thus, under the new directives, as previously, member states may continue to privilege considerations of public interest in judging alleged infringements of EC public procurement rules. In these respects, the mechanisms for ensuring conformity with the EC procurement rules can be said to remain 'weak': "The supranational objectives, encouraging decentralized control through the establishment of incentives, may not be attainable as the 'public interest' aspects which prevail in national legal approaches impose limits on the extent and scope of possible mechanisms" (Martin 1996: 229) . 97
7. Policy Implications: Future Prospects for GTP
It is evident that Government Technology Procurement is an innovation policy instrument. As such it is a matter of the interaction between procurers and suppliers, i.e., it is a matter of interaction between elements of the system of innovation. The point of departure is a perceived socio-economic problem or a socio-economic need that is not solved or mitigated by private market actors. Government Technology Procurement is, of course, only applicable when the problem/need can potentially be solved/satisfied through technical change. That is, for example, the case in the fields of energy, medical technology, defence, transport, communication and the environment.
However, Government Technology Procurement as an innovation policy instrument is poorly documented and therefore poorly understood by most policy-makers. There is a lot of ignorance around this subject -- for example, there is a persistent idea that procurement is associated with protectionism.
Government Technology Procurement is more difficult to practise the more fragmented the buyers are. The simplest case is when the procurer is a large government agency like the State Railway Company or the PTT. If the end users are single consumers the problems are the largest.
Government technology procurement requires forceful public actors who can identify needs, formulate demands, and thereby influence technical change. At the national level we have recently observed a process in which state agencies have been transformed into publicly controlled companies and these have sometimes been privatised. Such processes of decentralisation and privatisation make government technology procurement more difficult. Governments then no longer have a natural co-ordinator which can initiate procurement processes. The strength of the public sector is decreasing through the transformation of state agencies into companies and through privatisation of these. This process weakens the capacity to pursue a far-sighted and long-term government technology procurement policy with the dual objectives of solving societal problems and indirectly strengthening the competitiveness of (national) firms.
An efficient demand side innovation policy which includes government technology procurement requires a break in the trend toward decentralisation and privatisation, or an alternative strategy that would establish new procurement co-ordinating agencies or actors. One area where such an organization can be justified is in infrastructure for 98 information technology. In Sweden the PTT (Televerket) once had an over-arching responsibility in this field. It applied it in a visionary and well-administered fashion, as evidenced by the initiative to collaborate with Ericsson on the AXE electronic telephone exchange project or the formulation of the NMT 450 standard for mobile telephony. Though the need for long-range and visionary initiatives is even greater now, it is being neglected because no organization has been assigned the responsibility.
These examples demonstrate that user co-ordination and other organizational aspects of technology procurement are important components of a visionary demand-induced innovation policy. A specific conclusion, for the case of Sweden, is that some type of co-ordination organization with responsibility for a creative long-range strategy and pilot projects ought to be created as a substitute for earlier strong public actors. An attempt in this direction - in a small scale - is the procurement of energy efficient technology which was co-ordinated by NUTEK. Here it was a question of co-ordinating rather small and fragmented users. Individually they neither represented a large enough demand or enough competence to initiate a technology procurement process. NUTEK’s co-ordination was thus a necessary precondition for the procurement.
It is also often argued that government technology procurement is becoming more difficult (at the national level) because of the ongoing processes of internationalisation and European integration. We shall therefore conclude with a short discussion of this problem.
As a new member of the European Community (from January 1995), Swedish public procurement is now done in new forms. Every procurement deal is announced in the EC's Supplement to the Official Journal of the European Communities. Earlier Swedish laws for government procurement - i.e. demands for procurement on business lines, for competition and for non-discrimination - remain. As a matter of fact the differences in basic principles are not especially great when compared to the time before EU membership. The earlier Swedish rules did not constitute any obstacle for effective Government Technology Procurement and neither do the new ones. The rules with regard to Government Technology Procurement formulated by the EU and the WTO allow for the continuation of such activities (Westling 1996: ch. 12) . Hence Government Technology Procurement can continue to be used at the national level in countries that are members of the European Union.
A relevant policy question is, however, whether or not most European national systems of innovation are too small for handling procurement in some technological areas. (We might, for example, not need fifteen high-speed trains or aircraft designs in Europe but 99 perhaps only three or four.) If so, there might be arguments for using technology procurement more efficiently at the EU level in these areas. It might be a new policy option for the EU.
Technology procurement has the potential to serve as an innovation policy and industrial policy instrument for the Community as a whole. Earlier experiences at the national level should be of great value in an attempt to develop such a new policy option for the EU. The ISE procurement sub-project has the ambition to make such experiences available to policy makers.
In particular, it is important to investigate whether and in what ways technology procurement can be efficiently used as a policy instrument within the 5th Framework Programme. This is a relevant issue for the following reasons. The earlier framework programmes have been almost exclusively supply-push oriented. Various evaluations have indicated that this approach has not been very successful. This might be a reason why the preliminary discussion of the character of the 5th Framework Programme has a more problem-oriented character, seeking to change the largely technology-push emphasis of the past. If this, in the end, proves to be not only rhetorics, the new emphasis means that an availability of demand-pull oriented policy instruments is important. Government Technology Procurement is one of the few demand-side policy instruments available.
In the 'classical ' case of Government Technology Procurement, the Government uses nothing but its own market demand to induce technical innovation. It simply acts as a demanding buyer. This is the case in the defence sector (which is not studied in this sub- project) and in infrastructure (like telecommunications and trains) when government agencies are the final users of the technology or system. In other cases, the government acts as a co-ordinator and catalyst for user-induced technical innovation. It coordinates the demand and needs of other users. (This is the form of Government Technology Procurement practised by NUTEK in Sweden and addressed in section 2 above.)
Since the Commission does not represent final demand for goods, systems and services to any considerable extent, it is likely that the kind of procurement most relevant for the EU level is the one where EU organizations serve as catalysers and coordinate the demand for new products and services at the national level. In, for example, the cases of high-speed trains and civilian aircraft, it would be a matter of coordinating the efforts and requirements of a limited number of national and regional operating train and flight companies. 100
In a Systems of Innovation framework, it is natural to stress the demand side as well as demand side policy instruments. One of the most potent demand side instruments is Government Technology Procurement. It takes societal problems and needs as its point of departure. However, scientific and technological progress is a necessity to solve the problems and satisfy the needs being focused upon. It is natural to organise research projects as part of the study and project phases of technology procurement. Such research projects -which were discussed in section 2 above - could be an appropriate element of the activities of the EC.
The research projects should deal with questions about what is motivated from the needs point of view, what is socially desirable and whether it is technically possible to satisfy the needs identified within reasonable economic limits through technical development. In this way, results should be reached that can be a basis for the formulation of the functional demands in a technology procurement activity. The instrument of Government Technology Procurement could in this way serve as a co-ordinating mechanism between science and technology policy, innovation policy and industrial policy in the activities of the EC.
Part 3 of the sub-project, which will be carried out after the completion of the case studies, will draw policy implications from the sub-project's work as a whole. We expect to provide answers to, for instance, the following policy-relevant questions:
•What is the scope of using Government Technology Procurement at the regional, national, and European levels?
•Which kind(s) of GTP can fruitfully be used at these three levels?
•In what fields does GTP have the largest potential?
•Under what circumstances can Government Technology Procurement succeed?
•How should GTP be organised and managed in order, for example, to avoid problems of nepotism and corruption?
•How can the degree of competition between potential suppliers be increased and protectionism avoided?
•How can it be secured that GTP is practised as open bidding, that many potential suppliers are involved, and that the whole process is transparent? 101
•How can the capability of the procuring organisation with regard to specifying the functional requirements of the product or system be improved?
•What do the changes in EU procurement regulations following the Single Market mean for the national level and the European level?
•What other demand-side innovation policy instruments can be used?
•How can GTP be used in combination with other innovation policy instruments? 102
Appendix A. The EC Procurement Rules59
The present set of rules, or Regime, governing public procurement in Europe was formulated during the early 1990s as part of a broader effort to establish a single European market (Commission of the European Communities 1994) . The Regime's main objective is to promote "inter-EC procurement made without favouritism", thus helping to realize the broader purpose of "free movement of goods, persons, services and capital" (Business International Ltd. 1991) . The rules, therefore, express the underlying principles of giving procurement procedures a commercial character, maintaining competition and ensuring objectivity. "However, they (the rules) have been drawn up mainly for the purchasing of already developed goods or services, often standardized, and not specifically for development projects" (Westling 1996: Appendix 5, 1) .
A.1 Development of the Rules
Throughout the 1980s, the EC militated for the creation of a single 'internal market' for public purchasing within Europe, but made slow progress due to the lack of transparency in existing procedures and the lack of compliance on the part of member states with rules intended to remedy this problem. As late as 1989, it could be stated,
Of course the European Community has regulations for opening up public procurement to Community-wide competition, but not only are they incomplete, they are too often badly applied and even misused or ignored. Protectionist reflexes and deeply rooted chauvinistic attitudes are responsible for this state of affairs. (EC 1989)
Exasperation with this state of affairs prompted the Commission to launch a detailed investigation into the market for public procurement in Europe. The so-called Cecchini Report was delivered in 1988, and included among its main recommendations the following:
- making public purchasing procedures transparent, and open to all qualified bidders
59 * ' 1 1 < 103
- doing away with unnecessary differences in engineering standards and design specifications which confuse third-country customers as well as making intra-EC trade difficult
-viewing competition policy on an EC-wide basis, permitting both national and transnational mergers so that world-league EC firms can be built (Cecchini Report, cited inBusiness International Ltd. 1991: 6 - 7)
The consultancy research which provided the basis for these recommendations concentrated primarily on demonstrating "the costs of non-Europe" to the public sector (WS Atkins Management Consultants and Associates 1988) . It concluded that "The total estimated savings in public expenditure represent around half a percent of GDP, and this is equivalent to 2 to 5 percent of public purchasing, a very considerable saving" (WS Atkins Management Consultants and Associates 1988: 54) . The analysis and estimates described in this research projected that completion of the internal market could result in a potential saving of public sector expenditure in the range of 8 to 19 billion ECU, consisting of : ECU 3 - 8 billion in the five countries studied (Belgium, France, Germany, Italy, and the UK), arising directly from new trade at lowest-cost prices; a further ECU 1 - 3 billion resulting from competitive pressure on prices in sectors not previously open to competition; and, a further ECU 4 - 8 billion, arising eventually from restructuring leading to economies of scale in formerly protected sectors supplying equipment for defense, power generation, railways and telecommunications (WS Atkins Management Consultants and Associates 1988: 6 - 7) . In addition to these public sector savings, it was also expected that there were likely to be:
- benefits for private sector purchasers of similar goods.
- an important impact on the rate of innovation, investment and growth in the sectors enjoying the benefits of restructuring, with positive effects on their international competitiveness. (WS Atkins Management Consultants and Associates 1988: 7)
It is noteworthy that the previously protected sectors were a special target and focus of this research. It was stated, for instance, that the completion of the internal market would have few significant direct impacts on the bulk of producers in most countries, since -- among other reasons -- "the restructuring effects will be limited to certain key industries, for example, telecommunications equipment" (WS Atkins Management Consultants and Associates 1988: 6) . Elsewhere, attention was drawn to the fact that while existing EC Directives on transparency in public purchasing had had relatively 104 little success in "opening up" a broad range of sectors, this was especially the case in those sectors hitherto excluded from coverage by the directives: "This is ... particularly so for the purchasing sectors where there is really nationalistic purchasing for strategic reasons, and where there are large, specialist contracts of interest to foreign suppliers. ... They are: energy, transport, telecommunications, (and) drinking water supply. " (WS Atkins Management Consultants and Associates 1988: 12) Considerable effort was devoted to challenging the justifications for the continued exclusion of these sectors, and for this reason they were made a special focus of the research:
A special situation exists in the case studies and similar sectors where the public sector is the main purchaser, contracts are large, and there are few suppliers. Public purchasing in these sectors has in the past been used to maintain an indigenous capability for strategic reasons, and also at the same time to be an organ of anti-trust policy, to maintain more than one supplier. This had led to a situation in many sectors where there is virtually no trade between EC producers. Railway, power generation and telecommunications equipment are prime examples. Consequently, there is overcapacity, too many firms, duplication of R&D and marketing efforts, and suboptimal scale of production. In some of these sectors ... technology changes and the threat from new, low-cost producers have now made firms want to merge, rationalise, and invest in new technology. Wider public procurement is essential to permit this; without it, protectionist countries will be left with no viable producer. (WS Atkins Management Consultants and Associates 1988: 14 - 15)
This, then was the rationale informing the development of new EC rules to consolidate a single 'internal market' in Europe for public purchasing. The watchwords were transparency, rationalization of standards, and competition. The immediate objectives were to achieve public sector savings and to create dynamics leading to benefits for private sector purchasers. The longer-term objectives had to do with the restructuring of key sectors of European industry, in order to achieve international competitiveness.
A.2 What the Rules Cover
The current EC rules governing public procurement deal primarily with 'source selection' -- that is, the process in a public procurement transaction whereby the buyer finds a supplier. Very little is said about the public procurement contract itself, which is regarded as a private concern of the buyer and seller. The main emphasis of the rules is thus 'international competitive source selection' (ICS), the object being "to force EC covered entities to source competitively throughout the EC in procurements estimated by the buyer to be within the relevant threshold" (Business International Ltd. 1991: 15) . 105
The main initiating mechanism on which ICS relies is that of EC-wide publication (in the EC Official Journal ) of a notice of the proposed procurement, in a prescribed form. The main mechanisms for ensuring compliance and the development of appropriate remedies are suits brought by, or on behalf of, 'unhappy enterprises' to the European Court of Justice or lesser, domestic tribunals. (Both types of mechanism will be dealt with in later subsections.)
The rules can be said to be procedurally incomplete, both in the sense that they do not deal with the details of procurement contracts and in the sense that they do not deal with all aspects of source selection. With respect to source selection, the rules deal only with advertising, prequalification (the procedures an enterprise must follow before being allowed to submit a tender, by proving its standing and ability to fulfill the proposed procurement contract) and contract awards. The rules do not deal comprehensively or exhaustively with all aspects of listing (of invited or qualified candidates for procurement contracts) and tendering (the submission by candidates of formal proposals or procurement contracts) (Business International Ltd. 1991: 17) . Where the rules do not apply, the procurement process is covered by 'other' domestic law.
In substantive terms, the rules -- as of the mid-1990s -- now cover procurement by public utilities (formerly not covered) as well as public authorities (formally covered). They also now extend to services (not covered prior to 1992) as well as supplies and works (as previously). Figure 2 provides an overview of the internal market public procurement legislation with different Directives for for public authorities and utilities in respect of supplies, works and services.
Figure 2: Internal Market. Public Procurement legislation. Overview.
Public Authorities Supplies Works Services Remedies 77/62/EEC 80/767/EEC 71/305/EEC 88/295/EEC 89/440/EEC 92/50/EEC 89/655/EEC
93/36/EEC 93/37/EEC 106
Utilities Supplies, Works and Services Remedies
90/531/EEC 93/38/EEC 92/13/EEC
(Source:Westling 1996: Appendix 5, 1)
As mentioned previously, this legislation applies only above certain thresholds of estimated contract value. The rules are clear that no procurement may be split up with the intention of avoiding the threshold rules, and enterprises may complain of such practices to the Commission. The thresholds, expressed in ECUs, are intended to be revised periodically (Business International Ltd. 1991: 23) . Figure 3 provides an example of recently used threshold values.
Finally, note should be made of those areas in which the procurement rules do not apply. There is, for example, no application of the rules to items covered by Article 223.1(b) of the Treaty of Rome and maintained on the Commission's list of material "intended specifically for military purposes" -- (although the rules do apply to procurement of dual purpose materials) (Business International Ltd. 1991: 19) . Under Article 4 of Council Directive 93/36/EEC, which concerns international projects, other areas in which the rules do not apply include the following:
(a) in pursuance of an international agreement concluded in conformity with the Treaty, between a Member State and one or more non-member countries and covering supplies intended for the joint implementation or exploitation of a project by the signatory States; all agreements shall be communicated to the Commission, which may consult the Advisory Committee for Public Contracts set up by Decision 71/306/EEC;
(b) to undertakings in a Member State or a non-member country in pursuance of an international agreement relating to the stationing of troops;
(c) pursuant to the particular procedure of an international organization. (c.f. Westling 1996: Appendix 5, 4)
Figure 3. Threshold Values (figures in ECU net of VAT) 107
Works Supplies Services* Public Sector 5,000,000 200,000 200,000 (125, 672)** Utilities Sector: water, energy and 5,000,000 400,000 400,000 transport
Utilities Sector: telecommuni- 5,000,000 600,000 600,000 cations * Where a design contest is organised independently of a procedure for the award of a services contract, the relevant threshold for all prizes and payments is ECU 200,000. In the case of the Utilities sector, the relevant threshold will be the same as for contracts. ** Threshold value in the case of contracting authorities under GATT.
(Source: Trepte 1993)
Last, but not least, it should be noted that justifiable exceptions can be made to the rules in the case of development projects. For example, Article 8, paragraph 3 (93/36/EEC) of Title II Common Rules in the technical field states that a contracting authority may depart from the use of European standards if:
(d) the project concerned is of a genuinely innovative nature for which the use of existing European standards, European technical approvals or common technical specifications would not be appropriate. (c.f. Westling 1996: Appendix 5, 5)]
These exceptions to the rules will be further discussed in later subsections.
A.3 Rules for Source Selection -- With Reference to Development Projects
The rules prescribe three basic source selection procedures: the open procedure, the restricted procedure, and the negotiated procedure. Of these, it is the last two that are most applicable to the case of development projects of an innovative nature. However, it is also important to discuss the open procedure, since the spirit of the rules is to encourage its use in every possible case. 108
A.3.1 Open procedure
Following the open procedure, any interested supplier may submit a tender. The buyer must produce a notice of the proposed procurement in the prescribed form and send it to the EC's Office for Official Publications. There are strict criteria governing both preparation and publication of the notice. Generally, interested businesses have a period of 52 days, dating from despatch of the notice to the EC, to submit their tenders. Buyers are obligated to provide businesses applying with any contract and supporting documents they may require within four working days of receiving their applications. The buyer must also supply any additional information requested by applicants no later than six working days before the final deadline for receiving tenders. Where tenders can be made only after inspection of contract documents, the final deadline must be extended accordingly. Both buyers and interested businesses are free to use a variety of media to make invitations or requests for documents, but must give written confirmations. (Business International Ltd. 1991: 26 - 27)
Generally, the open procedure is meant to be applied in all cases involving the purchasing of already developed and standardized supplies, works and services. The use of technical standards is, therefore, a key issue in determining whether or not the procurement process is truly open. (This issue will be dealt with in more detail in a later subsection.) Generally, it is intended that the technical specifications made by the buyer should not stipulate special trade marks or processes that would exclude or favour certain prospective suppliers. However, the buyer also has a right, according to Article 16 of Title IV, Chapter I (common rules of participation) to make minimum technical specifications and should state if variants are not permitted, since the purpose of making such specifications is ideally to seek variants that improve upon them. Thus, it is allowable under the open procedure that the criterion for award of a contract can be 'the most economically advantageous offer', so that the contracting authorities may take account of variants which are submitted by a tenderer and meet the minimum specifications required" (Westling 1996: Appendix 5, 6) .
Technical standards are, of course, only one means of qualification, and in the open procedure, especially, they are meant to be used primarily in the evaluation of tenders, rather than in the prequalification of firms applying for a procurement contract. There are, however, other criteria which buyers are expected to apply in deciding whether or not enterprises should be excluded from 'participation in the contract', and these are meant to be applied not only in the open procedure but in others as well. An enterprise may be excluded if is insolvent, bankrupt, in an arrangement with its creditors, or subject to any related kinds of arrangements or legal proceedings. Exclusion may also 109 be justified if an enterprise has been either convicted or otherwise proven guilty of an offence concerning professional conduct. An enterprise may be excluded if it has not fulfilled its obligation to pay taxes, in either its home country or that of the buyer. Misrepresentation to the contracting authority of information concerning the enterprise's financial, economic or technical standing, or inability to prove that it is conducting business of the kind, in the location, and under the name specified in its submission to the contracting authority may also result in exclusion. Within this general frame of reference, the buyer may exclude an enterprise that fails to satisfy the buyer of its technical capacity:
Depending on the procurement, this can be done in a number of ways. A description of the enterprise's technical facilities, measures for quality control, and study and research facilities may be required whether the procurement is services ..., supplies or works. Similarly, the enterprise may be required to indicate educational and professional qualifications, membership of technical bodies, and quality control methods. And the buyer or a competent official body on its behalf may carry out an on-site check on the enterprise's production facilities, in the case of a complex or special procurement. (Business International Ltd. 1991: 31)
Once the correct tendering procedures have been completed, the contract must be awarded by following set criteria. Ideally, and especially in the open procedure, the primary criterion should be that of 'lowest price'. Alternatively, however, the buyer may decide on the 'economically most advantageous' tender or offer. In such cases, various other sub-criteria may be applied. These may include not only price, delivery date and cost-effectiveness criteria but also quality, aesthetic and functional characteristics, technical merit and post-sales service or technical assistance. If they are to be used, however, such criteria have to be specified in advance by the buyer in the notice and contract documents, and they can not be changed after the closing date for receiving tenders. Buyers are expected to take account of, and demand satisfactorily objective explanations for, abnormally low tenders. (Business International Ltd. 1991: 33)
A.3.2 Restricted procedure
Development projects and other procurements of an innovative nature are more likely to be dealt with under the restricted procedure. The restricted procedure differs from the open procedure primarily in that it relies on the selection or invitation of enterprises eligible to submit tenders. Under Article 19 of Council Directive 93/36/EEC, contracting authorities are allowed to "select from among the candidates with the qualifications required by Articles 20 to 29 those whom they will invite to submit a 110 tender or negotiate" (c.f. Westling 1996: Appendix 5, 7) . However, it is also required that the number of candidates invited to tender be sufficient to ensure competition. Although the contracting authorities can prescribe the range within which the number of suppliers they intend to invite will fall, it is stipulated that "The range must number at least 5 and may be up to 20" (c.f. Westling 1996: Appendix 5, 7) .
The justification required for use of the restricted procedure remains poorly defined, but the essential rationale is that "justification may be constituted by a need to maintain a balance between contract value and procedural costs, and by the specific nature of the goods to be procured" (Business International Ltd. 1991: 27) . The procedure has tended to be used as a matter of course in many areas of procurement, particularly services. However, contracting authorities are required to maintain a record of supplies contracts awarded using the restricted procedure, and forward details to the Commision should it request them to do so.
Public utilities operating in the water, energy, transport and telecommunications sectors have greater freedom in source selection and advertising than do public authorities. Article 20 of Council Directive 93/38/ EEC states that "Contracting entities may choose any of the procedures described ... provided that a call for competition has been made" through notice, periodic indicative notice, or notice on the existence of a prequalification system (Westling 1996: Appendix 5, 8) . Thus, utilities may set up prequalification systems according to prescribed procedures, and they may have have exclusive recourse to the resulting lists of prequalified businesses. They can also choose to advertise, not by means of a notice, but by either of two additional alternatives: a periodic indicative notice or a notice indicating the existence of a prequalification system. The last mentioned option requires no further notice of any specific proposed contracts, and the form of notice specified "does not require the utility to give details of any proposed procurement" (Business International Ltd. 1991: 29 - 30) . In these respects, public utilities have considerable latitude in the use of restricted procedures.
The use of restricted procedures by public authorities is more closely circumscribed and closer in form to the open procedure. The buyer is required to place a notice in the Official Journal describing the proposed procurement and inviting businesses that are interested to ask for further information. Interested businesses, or 'candidates', then normally have 37 days (from the date on which the buyer despatched the notice to the EC) to submit their requests for information -- in effect, their requests for invitation. 'Successful candidates' are then to be simultaneously invited by the buyer to submit tenders, together with letters of invitation, contract and supporting documents. These submissions must normally be made within 40 days of the date on which the invitations 111 were despatched. Enterprises' requests for further information must be met by the buyer within six working days of the final tender receipt deadline. The deadline itself can be extended in case tenders can only be made after inspection of contract documents. A variety of communications media can be used for invitations and requests, but written confirmation is required in all cases (Business International Ltd. 1991: 27) . The awarding of contracts follows the same general criteria as that described for the open procedure.
A.3.3 Negotiated procedure
The negotiated procedure is the most exclusive and is most likely to be associated with highly innovative development projects. Because it normally involves little or no competition, the negotiated procedure can be used only with adequate justification. Council Directive 93/36/EEC recognizes that "provision must be made for exceptional cases", but insists that "such cases must be expressly limited", so that "the negotiated procedure should be considered to be exceptional and therefore applicable only in limited cases" (c.f. Westling 1996: Appendix 5, 7) .
There are a number of conditions under which buyers may be permitted to use the negotiated procedure, and many of these involve requirements for innovation. The negotiated procedure can be used if no tenders or irregular tenders are received in response to open or restricted procedures. It can be used where the articles involved are manufactured only for research, experiment,study or development, though not produced in such quantity as to achieve commercial viability or cover R&D costs. The negotiated procedure can also be used for technical or artistic reasons, or to protect exclusive rights, when the goods can only be provided by a particular supplier. Additionally, its use can be justified in cases of extreme urgency, or where a change of suppliers would not be feasible for technical reasons. [ Business International Ltd., 1991 #309: 28]
Generally, the rules encourage continuity beween negotiated and restricted procedures. Negotiated procedures can be followed with or without publication, depending on the circumstances, but normally there are publication requirements to be followed in 'going over' to negotiated procedures. Article 6 (93/36/EEC) states:
... The contracting authorities shall in these cases publish a tender notice unless they include in such negotiated procedures all the enterprises satisfying the criteria of Articles 20 to 24 which, during the prior open or restricted procedure, have submitted tenders in accordance with the formal 112
requirements of the tendering procedure. (c.f. Westling 1996: Appendix 5, 7)
Public utilities, however, will generally have fewer such requirements to meet, for reasons explained in subsection 5.3.2. With respect to both negotiated and restricted procedures (with or without prequalification) for public utilities, paragraph 7(c), Article 1 (93/38/EEC) simply states that:
(c) In the case of negotiated procedures, the contracting entity consults suppliers, contractors, or service providers of its choice and negotiates the terms of the contract with one or more of them. (c.f. Westling 1996: Appendix 5, 8)
Generally, as with the restricted procedure, the contracting authority must draw up a written report justifying the use of negotiated procedures. This should provide a detailed description of the goods produced, the number of requests to participate received, the number of candidates invited to make an offer, and the number of candidates rejected, together with the reasons for rejection. Other circumstances leading to the use of the negotiated procedure should also be accounted for. The report must be provided to the Commission at its request. Beyond these requirements, however, the procurement process is subject to few, if any, additional procedural rules: "the buyer generally goes straight to its choice of supplier and negotiates the contract". (Business International Ltd. 1991: 28)
A.4 Rules Concerning Standards -- EC and International
In section 5.1 it was pointed out that a central policy objective of the Regime governing public procurement in the EC has been to eliminate unnecessary differences in engineering standards and design specifications. The rationale has been that such standards represent obstructions to the 'opening-up' of a single internal market for public purchasing, confusing third-country customers, and frustrating intra-EC trade. Thus, for example, it was noted in research on 'the costs of non-Europe' that "in certain key high-technology supplying sectors -- notably capital equipment for defence, power generation, telecommunications and railways -- a symbiotic relationship builds up between suppliers and purchasers" (WS Atkins Management Consultants and Associates 1988: 4) . The existence of such relationships was attributed in large part to a situation in which 113
There are widely differing national design standards, for example in railways and power generation. This creates a barrier to to the trade in goods and services supplied by the public sector, with widespread repercussions on transport and telecommunications, and hence on the efficiency of European business. (WS Atkins Management Consultants and Associates 1988: 4)
There are three basic approaches that the EC has taken in its efforts to resolve such problems through the rationalization of standards: the prevention of new barriers, the so called 'new approach' directives, and the development of new arrangements for testing and certification. The following discussion will outline these approaches and then proceed to consider those cases where a departure from established EC standards may be justified. Such cases, naturally, include development projects involving technological innovation.
A.4.1 Prevention of new barriers
Since 1984, EC member states have been required, under Directive 83/91/EEC, to inform the Commission of proposals for new technical standards. The directive now (since 1989) applies to all industrial and agricultural products. A moratorium of three months duration is imposed, allowing the Commission and member states to examine the contemplated standards and determine whether or not they are, in fact, a barrier to trade.
More generally, the rules attempt to prevent buyers from using technical specifications to discriminate in favour of their traditional suppliers. Such discrimination may often be covert. For this reason, buyers are as a rule obligated to express the relevant technical specifications in common terms. Common terms are usually either European standards, or national stangards which implement European standards intelligibly. There are, however, some justifiable departures from this rule. They include: cases where conformity with European standards cannot be established, either for technical reasons or because of deficiencies in the standards; certain cases of telecommunications equipment procurement; and, cases where using the prescribed standards would result in incompatibilities with materials already in use, or would result in excessive costs, or lead to technical difficulties.
Public works contracts are governed by special provisions regarding standards. These special provisions, however, illustrate the general approach to dealing with exceptional cases. Thus, in the case of public works contracts, it is allowed that technical 114 specifications can be expressed in terms of national standards for design, method of calculation, execution and use of materials. The specifications, however, must be related where possible to: national standards implementing international standards approved by the buyer's country; other national standards and technical approvals of the buyer's country; and any other technical standards. (Business International Ltd. 1991: 58 - 59)
A.4.2 'New approach' directives
The 'new approach' to harmonizing technical standards was initiated by the EC in 1985, through new directives that set out 'essential requirements' for products sold in EC countries. These requirements, which deal, for example, with health and safety and environmental concerns, are stated only in general terms, but they make reference to Eurpean Standards. The European Standards themselves are established by bodies mandated by the directives. These bodies are the European Committee for Standardization (CEN), the European Committee for Electrotechnical Standardization (Cenelec), and the European Telecommunications Standards Institute (ETSI), which also work on behalf of countries belonging to the European Free Trade Association (Efta). Pending complete harmonization of technical standards, CEN, Cenelec and ETSI award products whose manufacturers have demonstrated compliance with essential requirements the 'CE' mark, which permits them (the products) to be sold freely throughout the EC. (Business International Ltd. 1991: 54 - 55)
The international counterparts of CEN and Cenelec are, respectively, the International organization for Standards (ISO) and the International Electrotechnical Commission (IEC). These larger international bodies have established numerous product standards, which can be adopted by the corresponding European organizations as European Standards. Within CEN and Cenelec, standards are developed by technical committees with national representation and decision-making procedures requiring concensus between technical committees and national representatives. European Standards, whether they are adopted from ISO and IEC or developed independently, may be implemented as national standards within EC countries through national standards organizations. (Business International Ltd. 1991: 56) 115
A.4.3 Testing and certification
The 1989 Council Resolution on a Global Approach to Conformity Assessment (90/C/10/01) led to the establishment in 1990 of the European Organization for Testing and Certification (EOTC). This body facilitates mutual recognition of test results and certificates for non-regulated products among EC countries. It relies on European Standards to provide common criteria for assessing national certification bodies and test centers. Through this activity, EOTC seeks to minimize any unnecessary duplication of testing and certification. (Business International Ltd. 1991: 54 - 56)
As the 1989 Council Resolution on a Global Approach to Conformity Assessment (90/C/10/01), these provisions for certification and testing in the non-regulated sphere are an extension of the 'new approach' to the harmonization of technical standards. Guiding principles include, in addition to the promotion of mutual recognition agreements of certification and testing by a flexible, non-bureaucratic organization at the European level, a consistent approach in Community legislation for the various phases of conformity assessment procedures, encouraging member countries to set up compatible systems of accreditation using comparable techniques, measures to compensate for possible differences in levels of development among countries and industrial sectors, and efforts to promote trade in regulated products with third party countries through mutual recognition agreements. The Resolution also promotes generalized use of the European Standard for quality assurance (EN 29000) which adopts the ISO 9000 quality standard.
A.4.4 Departures from established EC standards
Some general cases where a justified departure from established EC standards is allowed in processes of public procurement have already been mentioned in section 5.4.1. Public works contracts are perhaps the most general of such cases, and they are therefore governed by special provisions which allow technical specifications to be expressed in terms of national standards while requiring, to the extent that this is possible, the national standards to be coordinated with relevant international, other national, and other technical standards. Other cases where a justified departure from established EC standards include cases where conformity with European Standards cannot be established, certain cases of telecommunications equipment procurement, and cases where using the prescribed standards would result in incompatibilities with materials in use, create technical difficulties, or incur excessive costs. 116
Generally, the rules impose more stringent conditions on buyers than on suppliers. Thus, under Council Directive 89/106/EEC suppliers preparing tenders are permitted to use different national technical specifications than those used in the buyer's country, so long as the technical specifications used can be defined with reference to European Standards; tenders cannot be rejected on these grounds. This principle is reiterated in Article 16, Paragraph 1 of Council Directive 93/36/EEC. Buyers, on the other hand, are enjoined to use European Standards and are only permitted to use alternatives to European Standards or other common technical specifications on condition that the reasons for doing so are documented and stated in the tender or contract documents. Buyers are also prohibited, as a matter of general principle, from specifying products of a make, source or process in a manner that would favourise or exclude certain EC suppliers or products. Buyers genuinely unable to describe the product they seek to procure in generic technical terms intelligible to a range of EC suppliers are permitted to refer in tender and contract documents to trademarks, patents, types, or product origins only if they also use the accompanying phrase "or equivalent".
Development projects, however, constitute an important exception to the rules regarding standards, which like the greater part of the EC procurement Regime, have been devised for application in the public purchasing of already existing goods that are, or should be, standardized. As observed in section 5.2, under Council Directive 93/36/EEC projects of a "genuinely innovative nature" in which the use of existing European Standards or common technical specifications "would not be appropriate" are exempt from these rules. Nevertheless, just as in cases where a departure is made from normal (open) tendering procedures, it is necessary to record the reasons for not utilizing European Standards. Paragraph 4, Article 8 of the same Directive (93/36/EEC) states:
... Contracting authorities ... shall record wherever possible thre reasons for doing so in the tender notice published in the Official Journal of the European Communities or in the contract documents and in all cases shall record these reasons in their internal documentation and shall supply such information on request to Member States and to the Commission. (c.f. Westling 1996: Appendix 5, 5)
Provided these conditions are met, the rules regarding standards present no hindrance to innovation. This is also the case for procurement by public utilities, which as pointed out in section 5.1 is governed by special provisions. In the Utilities Directive (93/38/EEC), Article 8, Paragraph 6 states that "contracting entities may derogate" if:
(d) the relevant European specification is inappropriate for the particular application or does not take account of technical developments which have come about since its adoption ... (c.f. Westling 1996: Appendix 5, 5) 117
International projects may also provide occasions justifying a departure from the EC procurement Regime, including its rules regarding standards. As observed in section 5.1, the rules do not apply to procurement processes carried out as part of international agreements involving countries which are not member states of the European Union. It is, however, required that these agreements be concluded in conformity with the Treaty of Rome and duly communicated to the Commission. Implicit in these requirements is the understanding that recognized international standards would be observed, through consultation with "the appropriate standards organizations" (Westling 1996: Appendix 5, 5) .
One sector in which the cases of innovative development projects and international agreements (possibly involving non-member states of the European Union) might coincide is the information technology (IT) sector. Arrangements for developing and applying appropriate standards in this rapidly evolving and strategically important sector have been described as follows:
EC wide standards are being developed by the Joint European Standards Institution (CEN/Cenelec), aided by the European Telecommunications Standards Institute (ETSI). Through ETSI, established in 1988, telecom network operators, users, research organizations, national telecom organizations and manufacturers are able to participate in the drawing up of new standards directly at European level. Bodies such as X/Open (an association of IT manufacurers using AT&T's Unix operating system) and the Open Systems Foundation (an association of IT manufacturers led by IBM) are assisting in the process. The regime specifically exempts the use of particular standards in telecommunications procurement if this would prejudice the application of Council Directive 86/361/EEC on the initial stage of the mutual recognition of type approval for telecommunications terminal equipment; or Council Decision 87/95/EEC on standardization in the field of information technology, which requires public sector buyers to specify Open Systems Interconnection (OSI) standards when purchasing IT systems above a certain cost threshold. (Business International Ltd. 1991: 105)
A.5 Mechanisms Ensuring Compliance 118
Prior to December 21, 1991, the EC's main mechanism for ensuring compliance with its Regime of public procurement rules was the administration of remedies under Article 169 EC by the European Court of Justice. Article 169 EC laid down an extensive set of procedures for ending alleged infringements by Member States of Treaty or secondary legislation. The Commission was (and remains) responsible for Article 169 EC's administrative implementation and the Court of Justice for its judicial application. Together these arrangements constituted a regulatory regime based on centralized supranational monitoring and systematic challenge of infringements that has been severely criticized for its "inadequacy" and "ineffective character". (Martin 1996: ch. 6) By 1987, the Commission itself was prepared to make the following statements about public procurement enforcement under Article 169 EC:
... the penalties it has been able to impose ... in response to complaints or on its own initiative had, in most cases, been imposed too late and were therefore counter-productive. Their ineffectiveness encourages infringements of the Community rules, and contracting authorities tend to regard the Commission's intervention under the procedure provided in Article 169 of the EEC Treaty as an arbitrary and unprovoked interference. Moreover, this ineffectiveness does not encourage contractors and suppliers to approach the Commission for assistance. (EC Commission, COM {87} cited in Martin 1996: 176)
For these reasons, the EC initiated, at the beginning of 1992, a transition towards more decentralized procedures for the supervising the application of Community Law. The rationale was to avoid "apoplexy at the centre and paralysis at the extremities" by achieving closer cooperation of Member States through the harmonization of national procedures for examining complaints (European Commission 1992) . Central to this more decentralised enforcement strategy initiated at the beginning of the 1990s was the development of adequate means of redress at the national level. This requirement was met by the so-called 'compliance' and 'remedies' directives discussed below.
A.5.1 Directives for compliance and remedies
Of the two main methods available to the Commission for the purpose of introducing new legislation -- Regulations and Directives -- directives are a more open-ended way of setting forth objectives to be achieved. They are "an imaginative way of first of all allowing a reasonable lead-in period before the introduction of new rquirements in complex areas such as procurement and secondly providing for different national characteristics to be reflected in the context of a European-wide piece of law" (Clarke and Temple 1994: 21) . Directives are thus a particularly appropriate method for the 119 development of a less centralized approach to ensuring compliance with public procurement rules.
Directives do not impose legislation directly on Member states but rather require only the 'harmonisation' of national laws with certain specific requirements, allowing this to be accomplished through a variety of forms and methods. At the same time, they clearly preserve important roles for the Commission and the European Court of Justice as central regulatory authorities. The Commission remains ultimately responsible for monitoring -- and enforcing, through infringement proceedings -- member states' compliance with any given directive. Moreover, "the question of whether or not a member state has successfully 'harmonised its laws, regulations and administrative provisions' ... with the terms of the directive can only be decided by the European Court of Justice and its judgements and rulings are definitive when trying to interpret the specific meaning of a particular requirement of a particular piece of legislation" (Clarke and Temple 1994: 22) .
The two main directives concerned with enforcement of the rules relating to public procurement are sometimes called the 'compliance' and 'remedies' directives, respectively. The first of these, Council Directive 89/665/ EEC, was promulgated to deal with infringements of the existing supplies and works directives -- 71/305/EEC (Works) and 77/62/EEC (Supplies). As of December 21, 1995, all member states were required to have this 'compliance' directive in place, for the purpose of providing effective application of the procurement rules missing in the earlier directives. The directive enables an unhappy business to challenge, in the buyer's country, buyer decisions that are perceived to be infringements of EC or related national law (Clarke and Temple 1994: 31) . The basic mechanisms for doing this have been described as follows:
Each member state must make available a tribunal (whether administrative or judicial in character) with power to grant certain minimum relief to an unhappy enterprise. The emphasis is on quick, effective intervention by the tribunal in the source selection decision-making process. Minimum relief available, in every such tribunal throughout the EC must include: (i) interim measures to suspend any part of the contract award procedure; (ii) setting aside any decisions taken unlawfully, including the removal of discriminatory technical, economic or financial specifications in the invitation to tender or any other document pertaining to the contract award procedure; (iii) damages to persons harmed by an infringement of EC law. (Business International Ltd. 1991: 37) 120
This directive was followed in 1990 by Council Directive 90/531/EEC, which subsequently became known as the 'Utilities Directive'. Rather than being concerned, as earlier legislation had been, with particular types of contracts, this directive referred to particular bodies placing contracts in what were formerly 'excluded sectors': water, energy, transport and telecommunications. The directive makes provision for the peculiarities of these sectors, and differentiates between works and supplies, but covers both types of contract in all four sectors. It also provides contracting authorities with considerable flexibility in regard to the form taken by calls for competition and free choice with respect to the options of open, restricted and negotiated procedures. (See previous discussion in sections 5.2 and 5.3). The Utilities Directive thus laid the foundation, not only for a subsequent directive concerning service contracts (93/38/EEC), but also for Council Directive 92/13/EEC, which concerned enforcement. The latter directive, also known as the 'remedies' directive, was brought into effect throughout the Community (except for now exhausted extensions in the cases of Spain, Greece and Portugal) in January, 1993. (Clarke and Temple 1994: 32 - 33)
Council Directive 92/13/EEC extended to the previously excluded sectors the same legal framework concerning enforcement of the procurement rules that had already been applied elsewhere in 89/665/EEC. There are, in fact, "few differences of significance between 92/13/EEC and 89/665/EEC":
In a similar way to its companion, for example, the Remedies Directive instructs Member States to set up an effective and speedy review procedure for contract awards falling within the scope (in this instance) of the Utilities Directive (90/531/EEC). As with the Works and Supplies Directives, the Utilities Directive contains no specific provisions to ensure the application of its rules.
Directive 92/13/EEC also allows for the possibility of conciliation at Community level in the event of disputes and calls for a substantial increase in the guarantees of transparency and non-discrimination. In a change from the Compliance Directive, however, the Remedies Directive provides for contracting authorities to have their procedures independently vetted under an attestation system which will allow those who comply with the rules to publicise this through the Official Journal.
Reverting then to the pattern of the Compliance Directive, there is a similar reservation by the Commission of its right to intervene directly when it sees a clear and obvious infringement of the award procedure. (Clarke and Temple 1994: 33)
A.5.2 Provisions 121
The 'compliance' and 'remedies' directives described above provide a flexible framework for ensuring the observance of the public procurement rules in which the onus of enforcement is placed "squarely on the business community" and, in addition to its having wide discretion with respect to interim measures, "all remedies are at the discretion of the (national) tribunal" (Business International Ltd. 1991: 37) . As explained previously, however, the European Court of Justice remains the final legal authority, and the European Commission remains free to intervene directly in clear cases of apparent infringement, as well as continuing to monitor more generally the application by member states of the directives. The argument has been made, therefore, that the modifications that have been made to the system of enforcement complete a development which "goes beyond a mere harmonization of existing national approaches" to define "an autonomous Community policy " in the area of public procurement:
This code ... overcomes national rules by the introduction of Community concepts and detailed award procedures. It also imposes autonomous principles and rules, such as the criteria for the award of the contract, which, moreover, have to be applied with reference to the interpretation given by the Court of Justice. ... This pre-emption of national competences has been completed by the extension of a public procurement regime to the area of enforcement level by the adoption of the Remedies Directive". (Martin 1996: 34 - 35)
Generally, the Regime depends on the actions of unhappy enterprises to invoke the procedures for remedying an infringement, through submission of a legitimate complaint to a national tribunal. The Regime requires the tribunals to conduct quick and effective reviews of source selection decisions for signs of contravention, and it also demands that they have consistent, reliable and harmonized intervention procedures. Nevertheless, member states have some latitude concerning how this is accomplished, and since reviews can only occur in the countries where a buyer's decision is being challenged, "the unhappy enterprise is dependent on the very jurisdiction it wishes to challenge" (Business International Ltd. 1991: 41) .
Basic provisions of the Regime (as summarized in Business International Ltd. 1991: 42 - 43) include the following:
(1) Member states may not discriminate between undertakings by exploiting differences between Community and national law. 122
(2) The remedies must be available to any person with an interest in a particular contract who is, or might be, harmed by an alleged infringement. Legal representation -- either by lawyers or other agents, such as trade associations -- and notice of intent to the buyer may be required of the person seeking the review. It is also open to individual members of a group to take such action on behalf of the group as a whole.
(3) There must be quick access to remedies and interim measures. Interim remedies may include the granting of injunctions, or the suspension of decisions concerning source selection or contract performance. In considering whether to award these, however, the tribunal may take into account not only private but also public interests, which buyers will usually identify themselves with.
(4) All member states have to provide for the suspension of contract awards, or implementation of other decisions made by buyers.
(5) Member states must also provide for the setting aside of unlawful decisions, including decisions made on the basis of discriminatory specifications in invitations to tender or dcuments related to contract award procedures.
(6) Member states must provide for the award of damages to persons harmed by infringements. Though there may be differences between countries in the categories of damages, these would generally include wasted tender costs, loss of expected profits, and loss of opportunities.
(7) The Commission is allowed to intervene directly in clear cases of infringement. The Commission can invoke its powers under EEC Treaty Article 169 to investigate and bring before the European Court of Justice failures to fulfill Treaty obligations on the part of member states.
In addition, there are some special provisions pertaining to infringements by utilities, reflecting the need for effective remedies at both the national and international level. Special features in the case of utilities (as summarized in Business International Ltd. 1991: 43 - 44) include the following:
(1) In claims for damages incurred in prparing tenders or participating in contract award procedures, applicants need only to prove that they were adversely affected by an infringement, but not to prove that the infringement resulted in the loss of a contract.
(2) Unhappy enterprises harmed or potentially harmed by an infringement may invoke a conciliation procedure by writing to the commission or to the responsible authorities in each member state. Both sides must pay thir own costs of participating in such a procedure. The procedure is overseen by an appropriate Advisory Committee of the Commission, which is responsible for convening a working group that requests representations from both sides and endeavours to achieve an agreement between them.
(3) Member states may institute "attestation" remedies as an alternative to making available interim measures such as suspensory injunctions. Attestation involves 123 compulsory annual auditing of utilities' purchasing procedures and practices in relation to domestic and EC law. The procedure is to be conducted by legally authorized and appointed independent reviewers experienced in procurement law and practice in the utility's field. The review bodies to whom their reports are made must be empowered to declare that infringements have occurred and to demand their correction or avoidance -- or failing these, that compensation be paid to the applicants.
A.5.3 Effectiveness
It is beyond the scope of this section and the article of which it forms a part, to offer a thorough analysis of the EC procurement rules, including those governing the Regime's enforcement. It is, however possible to draw attention to discrepancies between the original objectives of the directives concerning enforcement and the actual character of the directives.
As indicated in the preceding sections, the main conditions for effective remedies in the field of public procurement, as recognized in the EC procurement rules themselves, are: (i) fast review procedures allowing for early resolution of disputes, (ii) readily available interim measures permitting early resolution of disputes, and (iii) ensurance that adequate compensation will be paid to injured bidders in the event that early resolution is not achieved. The main purpose of of the directives concerning compliance and remedies was to correct insuffiencies with respect to all of these three conditions in existing national systems of redress.
It has been observed, however, that member states continue to enjoy considerable latitude with respect to all of the above-mentioned conditions. Thus, under the new directives, as previously, member states may continue to privilege considerations of public interest in judging alleged infringements of EC public procurement rules. In these respects, the mechanisms for ensuring conformity with the EC procurement rules can be said to remain "weak". The following assessment has been made of this state of affairs:
From a legal point of view, there was a need to strike a balance between the requirements of the effectiveness of redress mechanisms to achieve the aims of Community rules and the protection of the public and the private interests involved. The Commission's proposals disregarded public interest aspects and concentrated on effectiveness. An irreconcilable conflict between both aspects may exist. The supranational objectives, encouraging decentralized control through the establishment of incentives, may not be attainable as the 'public interest' aspects which prevail in national legal approaches impose limits on the extent and scope of possible mechanisms. (Martin 1996: 229) 124 125
Appendix B: Auctions as Procurement Mechanisms / by Martin Husz Why Bother about Auctions? At first sight both private and public procurement seem to be well defined problems of Neoclassical demand theory: Given your preferences, minimize procurement cost, that is, look for the lowest price of a good with a given quality. In a standardized market this theory may be applied in a relatively simple way. Imagine you want to buy a personal computer for multimedia applications. Before you buy it, you inform yourself reading through some computer journals (or find a friend who¥s playing instinct is better developed than yours and thus reads those journals anyway), then compare the prices in several shops and simply choose. The basic difference to procurement cases as discussed here is that there is no standardized market. Personal computers are sold continuously and to many costumers. In most of our procurement cases, however, the procuring party acts as a monopsonist and the good is usually procured only once (or at least in discrete points of time). There are neither journals on the procured item nor shops where several products can be compared. We are, furthermore, discussing products of innovative character. Thus, at the time when the procuring party decides on the procurement mechanism, the item to be procured is not yet fully determinded. Rather, the procuring party faces a more or less complete list of features the item should have, as for example in respect to its functionality or to the increase in technological competitiveness of national firms. A crucial aspect of the procurement problem is the asymmetry of information. The potential suppliers of the procured item have more information on the item than the procuring party. From the theoretical point of view, a procuring monopsonist faces an oligopoly on the supply side of the market, where the information of the suppliers is private. In such a situation the traditional market mechanism does not work well, as shown by Akerlof (1970). Suppliers can be expected to use their private information strategically, that is, they hide their information on production costs or on the quality of the good to be procured. Private information thus results in information rents. A second strategical opportunity for the suppliers is collusion: They can establish formal or informal cartels and cooperate against the procurer. The procurer, public or private, must design a pricing mechanism inducing potential suppliers to invent and provide the good at reasonable cost. As a monopsonist, the procurer has all the bargaining power because she has the advantage of the first move.60 Consequently, she can commit herself to a fixed pricing mechanism to induce competition between the suppliers. If suppliers cannot collude against the rules, then by this commitment the procurer may exploit all exchange surplus except the information rent. In most cases, a well designed auction is an optimal pricing mechanism for the procurer and, thus, minimizes the burden on tax payers in the case of public procurement. For that reason it makes sense to devote some attention to auction theory. The structure of this chapter is as follows: First, we define auctions and develop criteria to distinguish between different forms of auctions. Second, we describe some popular auction forms and show that they are equivalent under certain benchmark assumptions. Third, we relax these assumptions one by one and discuss why certain auctions might be superior to others. A short summary concludes the chapter. Definitions and Distinctions
60 In the words of economic theory: The procurer is a Stackelberg leader. 126
Auction models are partial equilibrium models adressing a subset of the class of adverse-selection problems as analyzed by Arrow (1985). Introducing auctions is equivalent to designing mechanisms, that is, processes that take bids as inputs and produce as output decisions on winners and payments. In defining an auction we follow McAfee and McMillan (1987, p.701): ``An auction is a market institution with an explicit set of rules determining resource allocation and prices on the basis of bids from the market participants.¥¥ This broad definition allows to discuss not only selling but also procuring by auction.61 Before we proceed by comparing several auctions we introduce criteria to distinguish and classify them. First, we discuss criteria which are inherent to the auction itself:
Bidding may be public (open-bid) or secret (sealed-bid). The winner(s) of the auction may be committed to the highest bid (first-price), to the second-highest bid (second-price) or to another rule determining the amount to be paid.62 The payment might also depend on variables other than the bids, as in the case of royalties, where the final price is paid only after revelation of the value of the auctioned item and relates to the latter. Auctions may have different stages of bidding until the winner(s) of the auction finally is (are) identified. We distinguish one-step auctions from auctions with recurrent bids. In the first case each bidder can only submit one bid whereas in the second case bids may be changed or repeated. An example for recurrent bids is the ascending-bid auction, where bids increase until the winner is identified. A more complex mechanism includes a change in the pricing criterion. Imagine a two-stage procurement mechanism where suppliers bid in terms of quality and then, in a second round, the winners bid again in terms of cost. The cost selection mechanism can also precede the quality selection mechanism and losers do not necessarily have to be excluded from the second round.
To compare auctions, however, we also must make assumptions on the bidders. These assumptions are crucial for certain results, which explains why we discuss classification criteria for the bidding parties: Bidders may be risk-neutral or risk-averse.63 Bidders may be independently different in their valuation of the auctioned item (private-values assumption) or may have correlated valuations (correlated-values assumption). As a special case of correlated values bidders may be identical in respect to their valuation of the good in question (common-value assumption). In the case of public procurement, the question is how much competeting suppliers differ in productivity and, hence, in project costs. Usually private values are assumed in models on public procurement. Correlated values seem more appropriate, however, if a common element of technological uncertainty is involved. This is likely to apply if public procurement is to encourage basic innovation.
61 Furthermore, the definition also implies double auctions, where several buyers and sellers submit bids simultaneously. Concentrating on public procurement, however, we will not further discuss double auctions here. 62 In the case of public procurement the highest bid is the lowest cost offered for a specific project. 63 We do not consider risk-seeking individuals here. Furthermore, we always assume that the auctioneer is risk-neutral. 127
From the viewpoint of the auctioneer bidders may be symmetric or fall into recognizably different classes. For example, in the case of public procurement asymmetric bidding occurs if local suppliers face a different cost structure compared to their competitors from abroad. Asymmetric bids imply correlated beliefs on the competitors¥ valuations. Classical Auctions and Benchmark Results Traditionally, four mechanisms out of a vast number of possible auctions are analyzed as benchmark cases. A first analysis has been executed by Vickrey (1961). These ``classical¥¥ auctions are: English Auction: The English auction is an open-bid, first-price, ascending-bid auction. Usually the auctioneer opens the auction by calling an initial price. The bidders successively raise their bids until the winner remains as the only bidder.64 The payment is equal to the winner¥s bid. Dutch Auction: The Dutch auction is an open-bid, first-price, descending-bid auction. The auctioneer calls an initial high price. The price is then succesively decreased until it is accepted by one of the bidders. The payment is equal to the bid. First-Price Sealed-Bid Auction: Each agent submits one sealed bid (one-step auction). The highest bid wins and the payment equals the bid. Vickrey Auction: The Vickrey auction is a second-price sealed-bid one-step auction. Each agent submits one sealed bid and the highest bid wins as in the first-price sealed-bid auction. Differently to the first-price sealed-bid auction, however, the payment is not equal to the amount given by the winning bid but to the amount determined by the second-highest bid. A first result concerns the Dutch auction and the first-price sealed-bid auction: The Dutch auction and the first-price sealed-bid auction are formally equivalent in the sense that they yield the same outcome.65 This result holds regardless of the assumptions on risk attitudes and value correlations. It is due to the equivalence of the bidder¥s decision problem: In both cases bidders cannot derive any information from the bidding in progress. Bidders depend on their ex-ante guesses of their opponents¥ strategies. They choose their highest bid given their beliefs on the highest bid of the others. Bidding ends in a Nash equilibrium where every bidder chooses a strategy determining the best answer on the anticipated moves of the competitors. The highest bid wins and determines the price of the good. In contrast to the English auction the decision of each bidder depends on the anticipated reactions of the competitors. If a bidder with a high valuation presumes that her opponents have a low valuation of the good, there is no reason to reveal her own preferences. Rather, this bidder will submit a bid only marginally higher than the expected second highest bid. Thus, the bidder is no price-taker as in perfect competition. Furthermore, beliefs can be erroneous. Though correct in average, the guess of the oppontents¥ valuation may be an underestimation in the actual case. Thus, a low-valuation bidder may win although another bidder would have placed a higher bid if her guess of the opponents¥ reactions would have been right. The situation is Pareto- efficient only on average.66
64 In the case of public procurement a ``high price¥¥ is equivalent to a cheap offer on behalf of the supplier. 65 For a proof see Vickrey (1961) 66 A situation is Pareto-efficient if no one can be made better off without making at least one other individual worse off. 128
Standard Assumptions: Before we proceed, we assume that the private-value assumption holds and that the payment is a function of the bids alone. Furthermore, we restrict the analysis to symmetric, risk neutral bidders. There are no collusions among the bidders. These assumptions are sufficient for a Pareto-efficient outcome of all the above auctions, at least on average. We relax these assumptions one by one in later chapters. In the English auction and the Vickrey auction the valuations of the bidders are revealed by the auction mechanism. In both cases the optimal strategy for a bidder is to bid according to her individual valuation. Note that this strategy is independent of the beliefs about the bidding of the competitors. We call such a strategy dominant. Dominant strategies result in a dominant equilibrium, which is a stronger equilibrium concept than the Nash equilibrium. In the case of the English auction, the revelation of the bidders¥ valuations follows from the fact that bidders keep bidding until their personal valuation is reached or until they win. Thus, the winner is the individual with the highest valuation of the good. The bidding stops when the bidder with the second-highest valuation drops out. Thus, the payment is equivalent to the second-highest bid as the winner would have aggreed on a higher price if necessary. In the case of the Vickrey auction, the bidders are price-takers as on a market with perfect competition. They have no possibility to influence the price on an individual level and, thus, have no incentive to hide their valuation. Lowering the bid would reduce the winning probability without increasing the potential payoff. Thus, the problem is similar to the English auction. Under the above assumptions the English auction and the Vickrey auction yield equivalent dominant equilibria.67 This result is only a special case of a more general finding, however. Under the above standard assumptions the English auction and the Vickrey auction are also quite similar to the Dutch auction and the first-price sealed-bid auction. In the latter two auctions, bidders base their bid on their beliefs about the strategies of their competitors. The final price of one of these auctions is the expectation of the second-highest valuation conditional on the winner¥s own valuation. Although this price only equals the second highest valuation by accident, it does so on average. Thus, on average the outcome of all four auctions is equal. Revenue Equivalence Theorem: Every auction allocating the goods efficiently and offering no profit to a zero valuation bidder has the same expected profits for every bidder valuation and the same expected revenue for the auctioneer.68 This is a nice theoretical result, yet contrasts empirical findings. English auctions are more frequent than other auctions. In certain environments like in private and public procurement problems, sealed-bid auctions dominate. One possible reason for this discrepancy between theory and evidence is that one or several of the above assumptions might not hold. We return to this point in the next sub-chapter, where we discuss which auction might be optimal for the auctioneer. Which Auction to Choose?
67 ibd. 68 For a proof see e.g. Milgrom (1989) p.10. 129
We say an auction is optimal if it maximizes the expected price for the auctioneer. In the case of public procurement this means that an optimal auction yields either the highest quality at a given cost or the lowest project cost for a given quality. From the Revelation Principle the outcome of any mechanism can also be reached by a direct, incentive compatible mechanism.69 Thus, we may limit our search for optimal mechanisms to the latter, where each bidder reports her valuation (rather than a price or another signal) of the item directly and where the mechanism is structured such that bidders have an incentive to reveal their true preferences. By applying the Revelation Principle some characteristics of optimal auctions can be derived.
Benchmark Case We start with our benchmark case, that is, we analyze the auctioneer¥s problem given the standard assumptions above. In this case the following result holds: If the standard assumptions hold, then an auction maximizes the expected price, (1) if the bidder with the highest valuation wins the auction (2) unless the resulting price falls short of a reserve price, in which case the auctioneer withdraws from the auction.70 According to this result, any of the English, Dutch, first-price sealed-bid, and Vickrey auctions is an optimal mechanism provided there is an optimally set reserve price. The finding is quite impressive because it implies that the simple auction forms are the best out of a large set of possible mechanisms. The reserve price is a powerful tool in the hand of the auctioneer. The intuition is similar as in the case of a textbook monopsony: From profit maximization the monopsonist tries to distort the outcome away from Pareto efficiency. In a world with asymmetric information a possible way to do so is by threatening to withdraw from the deal if the average winning bidder does not share some of her average information rent with the auctioneer. The auctioneer can commit herself credibly to such a strategy by introducing a reserve price. For example, if in an English auction the reserve price is higher than the expected second bid, then the expected profit of the auctioneer exceeds her profit of an auction without reserve price. This is quite relevant for public procurement because a reserve price rules out the possibility that the winning supplier charges a very high price only because she faces no serious competition. On the other hand, there is a positive probability that the reserve price leads to an inefficient outcome. Optimality requires that the reserve price be higher than the auctioneer¥s valuation of the good. If the reserve price exceeds the highest bid, however, then the auctioneer must withdraw from the deal even though her valuation of the good might be lower than the valuation of some bidders.
Risk-Averse Bidders We now relax one of the above assumptions, namely the assumption of risk- neutral bidders. Instead, we assume that bidders are risk-averse, that is, they prefer a certain payment to a fair lottery. In the case of auctions the bidder obtains nothing if she looses and a positive rent if she wins. If the auctioneer succeeds in increasing the risk of loosing then she can generate extra profits. In the case of the English auction this is not possible because ascending bidding implies that bidders can always increase their bid before loosing. In a Dutch auction, however, risk-aversion is similar to ``bad nerves¥¥:
69 For further information see e.g. Harris and Townsend (1985). 70 For a proof see e.g. Riley and Samuelson (1981) 130
Risk-averse bidders tend to accept the offer ``too early¥¥ which increases the expected profit of the auctioneer. Thus, the Revenue Equivalence Theorem no longer holds. Similarly, risk-averse bidders submit higher bids in a first-price sealed-bid auction to avoid the risk of loosing. If bidders are risk-averse, then the Dutch auction and the first-price sealed-bid auction yield higher profits for the auctioneer than the English auction and the Vickrey auction.71 To actually optimize the mechanism several tools can be used. One is to increase the riskiness of the auction by keeping the number of bidders secret. Another method to increase the pressure on the potential winner is to collect a bidding fee which decreases with the bid. Speculative behavior on behalf of a potential winner consequently becomes riskier because a low bid does not only loose with a higher probability but becomes expensive in itself. Formally, the trick is to increase the certainty-eqivalent of bidders with low valuation and to decrease the certainty-equivalent of bidders with high valuation.72
Asymmetric Bidders We already mentioned that bidders may fall into observably different groups. For example, in public procurement governments often favor local suppliers over foreign suppliers, that is, the monopsonist discriminates the price. Is there a rationale for this strategy? As it turns out, there actually might be, at least if domestic suppliers systematically face higher production costs which turns them into low valuation bidders. We assume that the distribution of valuations is identical for local and foreign firms, with the exception of a higher mean abroad. Imagine a first-price sealed-bid auction. Preferential treatment of local suppliers forces the bidders from the high- valuation class to bid more aggressively than they otherwise would have done. By an optimal mechanism the auctioneer can increase the price on average. There is a trade- off, however, because aiding domestic firms implies an increasing probability that such a firm might actually win the auction. This outcome would not only be Pareto- inefficient but also imply a relatively low profit for the auctioneer. To summarize the argument: Favoritism might be a good strategy if foreign firms have a comparative cost advantage, yet only to increase the competition for the potential winners. If, by accident, a less productive domestic firm actually wins (which should be unprobable on average) then this is clearly suboptimal. ``If the distributions of valuations are identical except for their means, then the class of bidders with the lower average valuation are favored in the optimal auction.¥¥73 Asymmetry also implies that the English auction and the first-price sealed-bid auction yield different outcomes, that is, the Revenue Equivalence Theorem does not hold. Imagine two classes (domestic and foreign) with three members each. A domestic bidder faces 2 domestic and 3 foreign competitors. Conversely, a foreign bidder faces 3 domestic and 2 foreign opponents. The English auction again yields the same Pareto- efficient outcome as in the benchmark case. In a Dutch auction, however, bidders must anticipate their competitors¥ valuations. Let us compare a domestic and a foreign firm with identical valuation. The estimated gap between the own valuation and the second- highest valuation of the good will be different even though the valuations of the two
71 ibd. 72 For further information on this topic see Mathews (1983). 73 This theorem is a quotation from McAfee and McMillan (1987), p.715. 131 firms are equal. Thus, the outcome must differ from the result in the English auction. Examples both for higher and lower prices can be found.74
Bidders with Correlated Beliefs Risk-aversion implies that the English auction yields less revenue to the auctioneer. Yet English auctions are quite popular. Milgrom and Weber (1982) have developed a model accounting for correlated values which explains why English auctions might be superior to first-price sealed-bid auctions in certain circumstances. To keep things simple we assume risk-neutrality throughout this chapter. To depict the role of correlated values we start with a simple special case, namely the common value assumption. For example, assume that all contractors to a public procurement project are equally capable and, thus, basically face the same production costs. There is, however, some uncertainty about these costs which forces the suppliers to make individual estimates on the actual production cost. For simplicity, we assume that the estimates are unbiased. Each estimate consists of the cost variable, which is random, plus an individual error term, where the estimation errors are independent. Even though the error terms are independent, the estimates are correlated because they share the common cost variable. On average, the individual estimation error is zero. The winning bid is above the average valuation, however, which implies that the winner¥s estimation error must be negative. In other words, the auction selects bidders who underestimate production cost. This phenomenon is called the ``winner¥s curse¥¥. Winning the auction is bad news to the winner in the sense that it signals that all competitors thought the item was worth less than the winner¥s bid. Rational bidders do not commit the same error over and over again. Thus, the winner¥s curse implies that risk-neutral bidders should also reduce their bid by a markup covering the expected risk from winning. Always being careful, is it still possible to win an auction? The answer to this question is perhaps disappointing: ``The most important lessons to be learned from both the theory and the experiments are that the returns in bidding come from cost and information advantages, /.../, and that bidders without some advantage have little hope of earning much profit, but could with a little bit of carelessness suffer large losses.¥¥75 In the case of public procurement we assume such ``cost and information differences¥¥ are always present. The argument of the winner¥s curse carries over to the more general case of correlated values where every bidder¥s guess on a competitor¥s valuation depends on her own valuation. To be more exact, following Milgrom and Weber (1982), we assume that the valuations of the bidders are affiliated. Valuations are affiliated if a bidder with a high valuation expects a higher valuation of a competitor than a bidder with a low valuation. Intuitively, imagine a potential supplier in a public procurement case informs herself and learns that the project might be more costly than predictable at first sight. Consequently, this supplier will usually assume that the competitors share the same information. Conversely, a supplier without such information will assume that other suppliers have as little information as herself. Thus, beliefs about other bidders¥ valuations are correlated to the own valuation. Affiliation Theorem: If valuations are affiliated, then the expected total surplus is the same in the English and the other ``classical¥¥ auctions, but the bidder¥s expected profit is smaller in the English auction and the auctioneer¥s expected
74 For further reading, see McAfee and McMillan (1987). 75 Milgrom (1989) p.6. 132
revenue is correspondingly higher. The auctioneer¥s return is second-highest in the Vickrey auction and lowest in the Dutch auction which yields the same revenue as the first-price sealed-bid auction.76 In the English auction bidders are price-takers, they reveal their true preferences by bidding. This argument does not change by affiliation of valuations. An ascending bid prevents the adverse effect of the winner¥s curse because a security against winning is no longer needed. In the Vickrey auction bidding is secret. The winner¥s curse thus reduces bids although preferences are revealed correctly. In the Dutch or first-price sealed-bid auctions, however, affiliation implies additional strategic speculations. Anticipation of the competitors¥ careful bidding induces the bidders to further reduce their bid. What can the auctioneer do to increase her profits in a sealed-bid auction? The answer is basically the same as in the case of an ascending bid: Additional common information decreases the adverse effect of the winner¥s curse und, thus, increases the auctioneer¥s revenue. This concords with Milgrom¥s informal Linkage Principle:77 The auctioneer is better off and the bidders are worse off if the winning price of the auction is linked to parameters affiliated with the bidder¥s private information. Royalties are one example for an application of the Linkage Principle. So far we have assumed that the payment resulting from the auction depends only on the bids. In contrast to this assumption, the auctioneer might condition payments on observable information about the winner¥s valuation. According to this argument, actually, royalties are a second example for price discrimination in auction theory. In the case of government procurement, the price paid by the government could be split into two parts where the first part depends on the bid and the second part depends on the ex-post observable success of the project. Suppose the project proves to be a great success for the winner of the auction who starts to sell the technologically advanced new product all over the world. The government could apply the Linkage Principle by participating in the contractor¥s success by a percentage tax on the product sold abroad. This must be known to the bidders ex-ante, of course. If the ``success¥¥-signal is distributed exogenously, then the auctioneer¥s revenue increases monotonically with the royalty rate.78 If the auctioneer¥s profit increases with the royalty rate, why should it be less than 100 percent? Theorists usually answer this question with a Moral-Hazard argument:79 If the contractor¥s cannot reap the revenue of her work, then she has no incentive to be successful, after all. If the royalty rate is too high, then the mechanism is not incentive-compatible and, thus, not optimal. Furthermore, the point on royalties can also be combined with the assumption of risk-aversion, where the intuition of the argument is quite similar as discussed in the respective chapter: The more risk-averse are the bidders, the higher should be the royalty rate.80
Collusive Bidders
76 For a proof see Milgrom and Weber (1982). 77 See e.g. Milgrom (1989) p.16. 78 See McAfee and McMillan (1986). 79 For example, McAfee and McMillan (1987) follow this line of argument. 80 ibd. 133
Collusions among the bidders are one key problem for the procuring party. By forming cartels bidders not only reduce the number of bidders and, hence, competition, they also change the effective distribution of valuations. A common counter-measure is an adequate choice of the reserve price. The anticartel reserve price is higher than the optimal reserve price in the absence of collusion. It increases with the number of cartel members.81 The auctioneer can also fight collusion itself. Cartels in repeated auctions work only if bidders departing from the group¥s agreement can be identified and punished by the other cartel members. In a sealed-bid auction this is only possible if the bids are publicly opened. If, in contrast, negotiations are closed as often found in the private sector, then there is an incentive to deviate from collusive arrangements.82
Variable Quantities In certain auctions, the auctioneer disposes of one more instrument to influence the auction¥s outcome: In industrial procurement the buyer can often vary the quantity bought. Rather than demanding a single item, the procurer¥s decision is represented by a demand function which correlates the quantity purchased to the price resulting from the auction. This is irrelevant if the auction is an English auction or a Vickrey auction, for bidders are then price takers. Suppliers then simply bid such that their marginal cost equals the offered price. This is not the social optimum, however. In the case of a first- price sealed-bid auction, the bidders face an additional strategic instrument. Suppose the suppliers¥ production function is such that they can offer at a lower cost provided they sell a higher quantity. By offering a contract at a lower cost they can increase their winning probability. This circumstance works to the auctioneer¥s advantage: The suppliers have an incentive to increase their bid, that is, to lower the project¥s price, in order to increase the contracted quantity. If the quantities are variable, then a first-bid auction leads to a higher expected quantity at a lower expected price (higher bid) than in a second-bid auction. The expected profit of both the auctioneer and (under relatively mild assumptions) the winning supplier are greater in an first-bid auction than in a second-bid auction.83 The intuition is relatively simple: Variable quantities allow for more competition in the first-bid auction. A higher quantity sold at a lower price implies a greater expected total surplus which can be shared by both sides of the market. For public procurement this implies that the auctioneer can improve the situation by auctioning off projects with price-dependent rather than fixed size. Note, furthermore, that a similar argument holds for quality, if it is possible to establish a monotonic relation between price and quality (e.g. price per quality unit). Conclusions In the framework of this book the basic question to be answered by auction theory is: How should a public procurer design an auction such that the technological competitiveness of her country can be improved at a low cost? Unfortunately, as in most academic analyses, the answer is: That depends. Yet, the analysis is probably helpful because it identifies the assumptions which allow clear recommendations. We always assume that suppliers are different in respect to their productivity.
81 ibd. 82 See e.g. Stigler (1964), p.48. 83 Hansen (1988) 134
In the benchmark case any traditional auction (English, Dutch, first-price sealed-bid, or Vickrey auction) can be chosen if a reserve price is set such that it maximizes competition. If we assume, furthermore, that bidders are risk-averse, then the auction should be a first-price sealed-bid auction. If domestic contractors have a lower productivity than foreign suppliers, then they should be favored in order to put more competitive pressure on the foreign contractors. Domestic suppliers should still have a low probability to win, however, as such an outcome would be suboptimal. If the beliefs of the bidders are correlated, then the procurer should provide as much public information on the project as possible. Although any other auction is better than the first-price sealed-bid auction under this assumption, the performance of the latter can be increased by contracting the procurement project as a royalty. If bidders tend to form cartels, the reserve price should be increased and the bids should be sealed and remain secret to the bidders. If the auctioneer has the possibility to vary the procured quantity or quality depending on the price, then she should do so in order to increase competition and total surplus and in order to minimize procurement costs.
Literature Akerlof George (1970):``The Market for `Lemons¥: Quality Uncertainty and the Market Mechanism.¥¥ Quarterly Journal of Economics 84(3), Aug., pp. 488-500 Arrow Kenneth J. (1985): ``The Economics of Agency.¥¥ In Principal and Agents: The Structure of Business. Ed. Pratt John W., Zeckhauser, Richard J., Cambridge, Mass.: Harvard Business School Press, pp. 37-51. Cassady Ralph Jr. (1967): ``Auctions and Auctioneering.¥¥ Berkeley: University of California Press. Hansen Robert (1988): ``Auctions with Endogenous Quantity.¥¥ Rand Journal of Economics 29(1), Spring, pp. 44-58 Harris Milton, Townsend Robert M.(1985): ``Allocation Mechanisms, Asymmetric Information, and the `Revelation Principle¥.¥¥ In Issues in Contenmporary Microeconomics and Welfare. Ed. Fiewel George, Albany: State U. of New York Press, pp.379-394. Matthews Steven A. (1983): ``Selling to Risk Averse Buyers with Unobservable Tastes.¥¥ Journal of Economic Theory 30(2), Aug., pp. 370-400. McAfee R. Preston, McMillan John (1986): ``Bidding for Contracts: A Principal-Agent Analysis.¥¥ Rand Journal of Economics 17(3), Autumn, pp. 326-338. McAfee R. Preston, McMillan John (1987): ``Auctions and Bidding¥¥; Journal of Economic Literature 15, June, pp. 699-738. Milgrom Paul (1989): ``Auctions and Bidding: A Primer.¥¥ Journal of Economic Perspectives 3(3), Summer, pp. 3-22. Milgrom Paul R., Weber Robert J. (1982): ``A Theory of Auctions and Competitive Bidding.¥¥ Econometrica 50(5), Sept., pp. 1089-1122. Riley John G. (1989): ``Expected revenue from Open and Sealed Bid Auctions.¥¥ Journal of Economic Perspectives 3(3), pp. 41-50. 135
Riley John G., Samuelson William F. (1981): ``Optimal Auctions.¥¥ American Economic Review 71(3), June, pp. 381-392. Stigler George J. (1964): ``A Theory of Oligopoly.¥¥ Journal of Political Economy 72(1), Feb., pp. 44-61 Vickrey William (1961): ``Counterspeculation, Auctions and Competitive Sealed Tenders.¥¥ Journal of Finance 16, March, pp. 8-37 136
References
Akerlof, G. 1970. The market for lemons: Quality uncertainty and the market mechanism. Quarterly Journal of Economics 84 (3):488 - 500. Allen, G. 1981. Industrial policy and innovation in Japan. In Industrial policy and innovation, edited by C. Carter. London: Heinemann. Allen, R. C. 1983. Collective invention. Journal of Economic Behavior and Organization 4 (1 - 24). Anchordoghy, M. 1988. Mastering the market: Japanese government targeting of the computer industry. International Organization 42:509 - 543. Andersen, E.-S. 1992. Approaching national systems of innovation. In National systems of innovation: Towards a theory of innovation and interactive learning, edited by B.-Å. Lundvall. London: Pinter Publishers. Arrow, K. 1962b. The economic implications of learning by doing. Review of Economic Studies XXIX (80). Arrow, K. 1974. The limits of organization. New York: W. W. Norton & Company. Arrow, K. J. 1985. The economics of agency. In Principal and agents: The structure of business, edited by J. W. Pratt and R. J. Zeckhauser. Cambridge, Massachusetts: Harvard Business School Press. Arthur, W. B. 1988. Competing technologies: An overview. In Technical change and economic theory, edited by G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete. London: Pinter Publishers. Baldwin, R. 1987. Preferential government purchasing policies. In Handbook of international economics, edited by K. Jones and J. Kenen. Amsterdam: Elsevier. Baldwin, W., and J. Scott. 1987. Market structure and technological change. London: Harwood Academic Publishers. Bannock, G., R. E. Baxter, and R. Rees. 1986. Penguin dictionary of economics. Third ed. Harmondsworth: Penguin Books. Barfield, C. 1994. Flat panel displays -- A second look. Issues in Science and Technology 1994 - 1995 (Winter):21 - 25. Bergström, E. 1980. Projectorientierad marknadsföring. Malmö: Liber Läromedel. Blume, S. 1981. Science policy research -- The state of the art and implications for policy., Report to the Swedish Council for Planning and Co-ordination of Research. Stockholm: Forskningsrådsnämden. Brookfield, H. 1975. Interdependent development. London: Methuen & Co. Burns, T., and G. M. Stalker. 1961. The management of innovation. London: Tavistock. 137
Burton, J. 1983. Picking losers? The political economy of industrial policy. London: Institute of Economic Affairs. Business International Ltd. 1991. Public procurement in Europe: How EC rules will open up new cross-border markets. London: Business International Ltd. Carlsson, B., ed. 1995. Technological systems and economic performance: The case of factory automation. Dordrecht: Kluwer Academic Publishers. Carlsson, B., and R. Stankiewicz. 1995. On the nature, function and composition of technological systems. In Technological systems and economic performance: The case of factory automation, edited by B. Carlsson. Dordrecht: Kluwer Academic Publishers. Choffray, J.-M., and G. Lilien. 1978. Assessing response to industrial marketing strategy. Journal of Marketing 42 (2):20 - 31. Clark, K. B. 1985. The interaction of design hierarchies and market concepts in technological evolution. Research Policy 14:235 - 251. Clarke, P. F., and E. M. F. Temple. 1994. Public procurement policy, Spicers European policy briefings. Harlow, Essex: Longman Group Ltd. Cohen, L. R., and R. G. Noll. 1991. The technology pork barrel. Washington, D.C.: The Brookings Institution. Cohen, W., and R. Levin. 1989. Empirical studies of innovation and market structure. In Handbook of industrial organization, edited by R. Schmalensee and R. Willing. Amsterdam: North Holland. Cohen, W. M., and D. A. Levinthal. 1989. Fortune favors the prepared firm. In Technical report. New York: Carnegie Mellon University. Cohen, W. M., and D. A. Levinthal. 1990. Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly 35 (March):128 - 152. Cohendet, P., and A. Lebeau. 1987. Choix stratégiques et grands programmes civils. Paris: Économica. Cohendet, P., and P. Llerena. 1997. Learning, technical change and public policy: How to exploit and create diversity. In Systems of innovation: Technologies, organizations and institutions, edited by C. Edquist. London: Pinter Publishers/Casell Academic. Commission of the European Communities. 1994. Public procurement in Europe: The directives. Brussels-Luxembourg: ECSC-EEC-EAEC. Cook, L., and J. Surrey. 1982. Government policy for the offshore supplies industry. Sussex: Science Policy Research Unit, University of Sussex. Cowan, R. 1995. The informatisation of government: From choice of technology to economic opportunity. STI Review (16):195 - 223. 138
Dahmén, E. 1970. Entrepreneurial activity and the development of Swedish industry, 1919-1939. Homewood: American Economic Association Translation Series. Dahmén, E. 1988. 'Development blocks' in industrial economics. Scandinavian Economic History Review 1. Dalpé, R. 1987. Politique d'achat et développement technologique. Quebec: Conseil de la science et de la technologie, Quebec. Dalpé, R. 1994. Effects of government procurement on industrial innovation. Technology in Society 16 (1):65 - 83. Dalpé, R., and C. DeBresson. 1988. Le secteur public comme premier utilisateur d'innovations. Montréal: Centre for Research on Development of Industry and Technology. De Bresson, C., and F. Amesse. 1991. Networks of innovators: A review and introduction to the issue. Research Policy 20:363 - 379. Dertoutzos, M. L. , R. K. Lester, and R. M. Solow. 1989. Made in America. Cambridge, Massachusetts: MIT Press. Dosi, G. 1982. Technological paradigms and technological trajectories. Research Policy 11:147 - 163. Eads, G., and R. R. Nelson. 1971. Government support of advanced civilian technology: Power reactors and the supersonic transport. Public Policy. EC. 1989. European file: opening up public procurement in the European Community. EC Official Journal (January-February). Edquist, C. 1994a. Technology policy: The interaction between governments and markets. In Technology policy: Towards an integration of social and ecological concerns, edited by G. Aichholzer and G. Schienstock. New York: Walter De Gruyter. Edquist, C. 1996a. Government technology procurement as an instrument of technology policy. In Technological infrastructure policy, edited by M. Teubal and Associates. Netherlands: Kluwer Academic Publishers. Edquist, C. 1996b. Policy implications of the Systems of Innovation approach. Paper read at session on 'New Concepts and Frameworks for Devising Policy' at the International Colloquium, 'RTD and Innovation Policies for Future Balanced Growth in Europe -- A Systems Approach', organized by FORFAS, 4 November, at Dublin. Edquist, C. 1997a. Introduction. In Systems of innovation: Technologies, organizations and institutions, edited by C. Edquist. London: Pinter Publishers/Casell Academic. Edquist, C., L. Hommen, and M. McKelvey. 1997. Innovations and employment in a systems of innovation perspective: ISE (Innovation Systems and European 139
Integration) report -- Sub-project 3.1.2: 'Innovations, growth and employment'. Brussels: EC. Edquist, C., and B. Johnson. 1997. Institutions and organizations in systems of innovation. In Systems of innovation: Technologies, organizations and institutions, edited by C. Edquist. London: Pinter Publishers/Casell Academic. Edquist, C.., and B.-Å. Lundvall. 1993. Comparing the Danish and Swedish national systems of innovation. In National systems of innovation: A comparative study, edited by R. Nelson. Oxford: Oxford University Press. Ergas, H. 1987. The importance of technology policy. In Economic policy and technological performance, edited by S. P. and P. Dasgupta. Cambridge: Cambridge University press. European Commission. 1992. Communication to the Council and European Parliament on the principle of subsidiarity. European Documents 1804 (05):14. European Commission. 1994. Public procurement in Europe: The directives. Luxembourg: Commission of the European Communities. Faucher, P. 1990. Large scale projects and technological development: An institutional approach. Paper read at Annual Conference of the Canadian Political Science Association, 27 May, at Victoria. Faucher, P., and K. Fitzgibbons. 1993. Public demand and the management of technological risk in large-scale projects. Science and public Policy 20 (3):173 - 185. Finsinger, G. 1988. Non-competitive and protectionist government purchasing behaviour. European Economic Review 32. Flamm, K. 1987. Targeting the computer: Government support and international competition. Washington, D.C.: The Brookings Institution. Flamm, K. 1988. Creating the computer. Washington, D.C.: The Brookings Institution. Folster, S. 1989. The efficiency of R&D subsidies. Stockholm: Industrial Institute for Social and Economic Research. Foray, D. 1989. Les modeles de compétition technologique: Un revue de la littérature. Revue d'Économie Industrielle 48. Foray, D. 1997. Generation and distribution of technological knowledge: Incentive structures, norms and institutions. In Systems of innovation: Technologies, organizations and institutions, edited by C. Edquist. London: Pinter Publishers/Casell Academic. Freeman, C. 1982. The economics of industrial innovation. Third ed. London: Frances Pinter. 140
Freeman, C. 1988. Japan: a new national system of innovation? In Technical change and economic theory, edited by G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete. London: Pinter Publishers. Freeman, C. 1991. Networks of innovators: A synthesis of research issues. Research Policy 20 (5):499 - 514. Freeman, C., and C. Perez. 1988. Structural crises of adjustment: Business cycles and investment behaviour. In Technical change and economic theory, edited by G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete. London: Pinter Publishers. Fridlund, M. 1993. The 'development pair' as a link between systems growth and industrial innovation: Cooperation between the Swedish State Power Board and the ASEA company. Stockholm: Royal Institute of Technology, Department of History of Science and Technology. Gelsing, L. 1992. Innovation and the development of industrial networks. In National systems of innovation: Towards a theory of innovation and interactive learning, edited by B.-Å. Lundvall. London: Pinter Publishers. Geroski, P. A. 1990. Procurement policy as a tool of industrial policy. International Review of Applied Economics 4 (2):182 - 198. Gibbons, M., C. Limoges, H. Nowotny, S. Schwartz, P. Scott, and M. Trow. 1994. The new production of knowledge: The dynamics of science and research in contemporary societies. London: Sage. Glassman, J. 1993. NAFTA and trade are tiny parts of a job revolution. The Washington Post, October 8, G1. Goodman, A., and W. Saunders. 1985. US federal regulation of foreign involvement in aviation: Government procurement and national security. Journal of World Trade 54. Graham, M. 1988. R&D and competition in England and the United States. Business History Review 62:261 - 285. Granstrand, O. 1984. Technology procurement as a special form of buyer-seller interaction in industrial markets: CIM report no. 84:06. Göteborg: Chalmers University of Technology, Department of Industrial Management. Granstrand, O., and S. Jacobsson. 1983. Innovation economy and business development: A study of innovation take-overs: CIM report no. 83:08. Göteborg: Chalmers University of Technology, Department of Industrial Management. Granstrand, O., and J. Sigurdson, eds. 1985. Technological innovation and industrial development in telecommunications: The role of public buying in the telecommunication sector in the Nordic countries. Lund: Research Policy Institute. 141
Grant, B., and L.-E. Gadde. 1984. Automotive component supply strategies: CIM report no. 84:03. Göteborg: Chalmers University of Technology, Department of Industrial Management. Grübler, A. 1989. Diffusion: Long term patterns and discontinuities. Paper read at International Conference on Diffusion of Technologies and Social Behavior, 14 - 16 June, at Laxenburg. Ham, R. M., and D. C. Mowery. 1995. Enduring dilemmas in US technology policy and the Clinton administration. California Management Review. Hansen, R. 1988. Auctions with endogenous quality. Rand Journal of Economics 29 (1):44 - 58. Harris, M., and R. M. Townsend. 1985. Allocation mechanisms, asymmetric information and the 'revelation principle'. In Issues in contemporary micro- economics and welfare, edited by G. Fiewel. Albany: State University of New York Press. Hartley, K. 1987. Public procurement and competitiveness: A Community market for military hardware and technology? Journal of Common Market Studies XXV (3):237 - 247. Hayek, F. A. 1948. Economics and knowledge and the use of knowledge in society. In Individualism and the economic order, edited by F. A. Hayek. Chicago: University of Chicago Press. Hellman, J.-L. 1993. Supplier-customer interaction in product development: A case study of Nippon Steel's cooperation with Toyota. Kunitachi, Tokyo: Hitotsubashi University, Institute of Business Research. Hidefjäll, P. 1997. The pace of innovation: Patterns of innovation in the cardiac pacemaker industry. Ph.D. Dissertation, Department of Technology and Social Change, Linköping University, Linköping, Sweden. Hodgson, G. 1991. Evolution and intention in evolutionary theory. In Evolutionary theories of economic and technological change: Present status and future prospects, edited by P. Saviotti and S. Metcalfe. Reading: Harwood. Holmes, J. 1986. The organization and locational structure of production subcontracting. In Production, work, territory: The geographical anatomy of industrial capitalism, edited by A. J. Scott and M. Storper. Boston: Allen & Unwin. Håkansson, H., ed. 1987. Industrial technological development: A network approach. London: Croom Helm. Håkansson, H. 1989. Corporate technological behaviour: Cooperation and networks. London: Routledge. Håkansson, H. 1990. Technological collaboration in industrial networks. European Management Journal 8 (3):371 - 379. 142
Håkansson, H., and H. Lundgren. 1993. Industrial networks and technological innovation. Uppsala/Stockholm: Network for Industry and Market Studies. Imai, K., and Y. Baba. 1989. Systemic innovation and cross-border networks: Transcending markets and hierarchies to create a new techno-economic system. Paris: OECD, Conference on Science, Technology and Economic Growth. Jeanrenaud, F. 1984. Marchés publics et politique économique. Annales de l'Économie Publique, Sociale et Coopérative 72 (2). Jeanrenaud, F., and L. Meyer. 1984. L'incidence a moyen et longe terme des commandes de matériel de télecommunication: Deux études de cas. Annales de l'Économie Publique, Sociale et Coopérative 72 (2). Johansson, B. 1991. Economic networks and self-organization. In Economic networks, innovation, and local development in industrialized countries, edited by E. M. Bergman, G. Maier and F. Tödtling. London: Mansell. Johnson, B. 1992. Institutional learning. In National systems of innovation: Towards a theory of innovation and interactive learning, edited by B.-Å. Lundvall. London: Pinter Publishers. Kamien, M., and N. Schwartz. 1982. Market structure and innovation. Cambridge: Cambridge University Press. Kanz, J. W. 1993. Technology, globalization and defence: Military electronics strategies in a changing world. International Journal of Technology Management 8 (1/2):59 - 76. Katz, B. G., and A. Phillips. 1982. The semiconductor industry. In Government and technical progress -- A cross-industry analysis, edited by R. R. Nelson. New York: Pergamon Press. Keyzer, M. 1968. L'importance et la politique des achats publics. Hommes et Techniques (278). Kline, S. J., and N. Rosenberg. 1986. An overview of innovation. In The positive sum strategy: Harnessing technology for economic growth, edited by R. Landau and N. Rosenberg. Washington, D. C.: National Academy Press. Levin, R. C. 1982. The semiconductor industry. In Government and technical progress - - A cross-industry analysis, edited by R. R. Nelson. New York: Pergamon Press. Levitt, B., and J. G. March. 1988. Organizational learning. Annual Review of Sociology 14:319 - 340. Lichtenberg, F. 1988. The private R&D investment response to federal design and technical competitions. American Economic Review 78:550 - 559. Linder, G. 1982. Teknikupphandling -- Något om förusättningar och problem. Stockholm: IVA. 143
Lundvall, B.-A. 1988. Innovation as an interactive process: From user-producer interaction to the national system of innovation. In Technical change and economic theory, edited by G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete. London: Pinter Publishers. Lundvall, B.-Å. 1985. Product innovation and user-producer interaction, Industrial development research series no. 31. Aalborg: Aalborg University Press. Lundvall, B.-Å., ed. 1992. National systems of innovation: Towards a theory of innovation and interactive learning. London: Pinter Publishers. Lundvall, B.-Å., N. Maarbjerg-Olesen, and I. Aaen. 1983. Det lansbrugsindustrielle kompleks: Teknologiudvikling, konkurrenceevne og beskæftigelse. Aalborg: Aalborg University Press. Maddock, I. 1983. Civilian exploitation of defence technology. London: National Economic Development Office. Mansfield, E., and J. Rapoport. 1977. Social and private rates of return from industrial innovations. Quarterly Journal of Economics (May):221 - 240. Martin, J. F. 1996. The EC public procurement rules: A critical analysis. Oxford: Clarendon Press. Matthews, J. 1996. Organizational foundations of the knowledge-based economy. In Employment and growth in the knowledge-based economy, edited by D. Foray and B.-Å. Lundvall. Paris: OECD. Mattson, L.-G. 1978. Systemförsäljning. Stockholm: Marknadstekniskt Centrum. McAfee, R. P., and J. McMillan. 1986. Bidding for contracts: A principal-agent analysis. Rand Journal of Economics 17 (3):326 - 338. McAfee, R. P., and J. McMillan. 1987. Auctions and bidding. Journal of Economic Literature 15 (June):699 - 738. McCrudden, C. 1994. Public procurement and equal opportunities in the European Community: A study of 'contract compliance' in the member states in the European Community and under Community law. Brussels: EC. McKelvey, M. 1997. Delineating evolutionary systems of innovation. In Systems of innovation: Technologies, organizations and institutions, edited by C. Edquist. London: Pinter Publishers/Casell Academic. McKinsey and Company. 1988. Strengthening competitiveness in UK electronics. London: National Economic Development Office. McLachlan, L. 1985. Discriminatory public procurement, economic integration, and the role of bureaucracy. Journal of Common Market Studies 23. Milgrom, P. 1989. Auctions and bidding: A primer. Journal of Economic Perspectives 3 (3):3 - 22. Milgrom, P., and R. J. Weber. 1982. A theory of auctions and competitive bidding. Econometrica 50 (5):1089 - 1122. 144
Mowery, D., and N. Rosenberg. 1979. The influence of market demand upon innovation: A critical review of some recent empirical studies. Research Policy 8:102 - 153. Mowery, D.C. 1995a. US postwar technology policy and the creation of new industries. Paper read at Conference on Creativity, Innovation and Job-Creation, January 11-12, at Oslo, Norway. Mowery, D. C., and N. Rosenberg. 1982. The commercial aircraft industry. In Government and technical progress -- A cross-industry analysis, edited by R. R. Nelson. New York: Pergamon Press. Mowery, D. C., and N. Rosenberg. 1993. The US innovation system. In National systems of innovation: A comparative study, edited by R. Nelson. Oxford: Oxford University Press. Nelson, R. 1981. Research on productivity growth and productivity differences: Dead ends and new departures. Journal of Economic Literature 19 (September):1029 - 1064. Nelson, R. 1987. Understanding technical change as an evolutionary process. Amsterdam: Elsevier. Nelson, R. 1988. Institutions supporting technical change in the United States. In Technical change and economic theory, edited by G. Dosi, C. Freeman, R. Nelson, G. Silverberg and L. Soete. London: Pinter Publishers. Nelson, R., ed. 1993. National systems of innovation: A comparative study. Oxford: Oxford University Press. Nelson, R. 1995a. Recent evolutionary theorizing about economic change. Journal of Economic Literature 33 (March):48 - 90. Nelson, R., and N. Rosenberg. 1993a. Technical innovation and national systems -- Introduction. In National systems of innovation: A comparative study, edited by R. Nelson. Oxford: Oxford University Press. Nelson, R., and N. Rosenberg. 1993b. In search of a useful theory of innovation. In National systems of innovation: A comparative study, edited by R. Nelson. Oxford: Oxford University Press. Nelson, R., and S. Winter. 1977. In search of a useful theory of innovation. Research Policy 6 (1):36 - 76. Nelson, R., and S. Winter. 1982. An evolutionary theory of economic change. Boston, Massachusetts: The Bellknap Press of Harvard University Press. Nelson, R. R. 1984. High-technology policies: A five-nation comparison. Washington, D.C.: American Enterprise Institute. Nielsen, K. 1991. Towards a flexible future -- Theories and politics. In The politics of flexibility -- Restructuring state and industry in Britain, Germany and 145
Scandinavia, edited by B. Jessop, H. Kastendiek, K. Nielsen and O. K. Pedersen. Aldershot: Edward Elgar. Nilson, H. 1994. Technology procurement for market transformation -- a customer tool for sustainable design and manufacturing. Stockholm: NUTEK. Niosi, J., P. Saviotti, B. Bello, and M. Crow. 1993. National systems of innovation: In search of a workable concept. Technology in Society 15:207 - 227. OECD. 1991a. Information technology standards: The economic dimension. Paris: OECD. OECD. 1996a. Technology, productivity and job creation -- Vol.1: Highlights, The OECD jobs strategy. Paris: OECD. OECD. 1996b. Technology, productivity and job creation -- Vol.2: Analytical report, The OECD jobs strategy. Paris: OECD. OECD/ICCP. 1996. Mobile cellular communications: Pricing strategies and competition, ICCP report, no. 39. Paris: OECD. Parker, P. 1990. The 1988 local government act and compulsory competitive tendering. Urban Studies 5. Parkinson, S. T. 1982. The role of user in successful new product development. R&D Management 12 (3):123 - 131. Parsons, H. 1988. Economic principles in the private and public sectors. Policy and Politics 16. Pavitt, K. 1984. Sectoral patterns of technical change: Towards a taxonomy and a theory. Research Policy 13:343 - 373. Perroux, C. 1969a. Note sur la notion de la pole de croissance. In L'economie du XXé siecle, edited by C. Perroux. Paris: Presses Universitaires de France. Perroux, C. 1969b. Les espaces economiques. In L'economie du XXé siecle, edited by C. Perroux. Paris: Presses Universitaires de France. Perrow, C. 1993. Small firm networks. In Institutional change: Theory and empirical findings, edited by S.-E. Sjöstrand. London: M.E. Sharpe. Phillips, A. 1980. Organizational factors in R&D and technological change: Market failure considerations. In Research, development and technological innovation, edited by D. Sahal. Lexington, Massachusetts: Lexington Books. Ponssard, J. P., and G. De Pouvourville. 1982. Marchés publiques et politique industrielle. Paris: Économica. Porter, M. E. 1990. The competitive advantage of nations. London: MacMillan. Powell, W.W. 1990. Neither market nor hierarchy: Network forms of organization. In Research in organizational behavior -- Volume 10, edited by B. M. Straw and L. L. Cummings. Greenwich: JAI Press. 146
Rankine, L. J. 1995. The role of users in information technology standardisation. STI Review (16):177 - 194. Riley, J. G., and W. F. Samuelson. 1981. Optimal auctions. American Economic Review 71 (3):381 - 392. Robinson, P., C. Faris, and Y. Wind. 1967. Industrial buying and creative marketing. London: Allyn & Bacon Ltd. and The Marketing Science Institute. Roll, L. 1982. The mixed economy. London: MacMillan. Rosenberg, N. 1972. Factors affecting the diffusion of technology. Explorations in Economic History 10. Rosenberg, N. 1976. Perspectives on technology. Cambridge: Cambridge University Press. Rosenberg, N. 1982. Inside the black box: Technology and economics. New York: Cambridge University Press. Rothwell, R. 1994. Issues in user-producer relations in the innovation process: The role of government. International Journal of Technology Management 9 (5/6/7):629 - 649. Rothwell, R., and W. Zegveld. 1982. Industrial innovation and public policy. London: Frances Pinter. Rothwell, R., and W. Zegveld. 1985. Reindustrialization and technology. London: Longmans. Sahal, D. 1981. Patterns of technological innovation. New York: Addison Wesley. Sahal, D. 1985. Technology guide-posts and innovation avenues. Research Policy 14:61 - 82. Saviotti, P. 1997. Innovation systems and evolutionary theories. In Systems of innovation: Technologies, organizations and institutions, edited by C. Edquist. London: Pinter Publishers/Casell Academic. Scherer, F.M. 1980. Industrial market structure and economic performance. Second ed. Chicago: Rand McNally. Schmidt, S. 1992. Negotiating technical change through standards: Technical coordination in market and committees. Cologne: Max Planck Institute. Sheth, J. 1973. A model of industrial buyer behavior. Journal of Marketing 37 (4):50 - 56. Shulman, P. R. 1980. Large scale policy-making. Elsevier: New York. Silverberg, G. 1987. Technical progress, capital accumulation, and effective demand: A self-organization model. In Economic evolution and structural adjustment, edited by D. Batten, J. Casti and B. Johansson. New York: Springer-Verlag. Silverberg, G. 1990a. Adoption and diffusion as a collective evolutionary process. In New explorations in the economics of technological change, edited by C. Freeman and L. Soete. London: Pinter Publishers. 147
Silverberg, G. 1990b. Dynamic vintage models with neo-Keynesian features. In Technology and productivity: The challenge for economic productivity, edited by OECD. Paris: OECD. Silverberg, G., G. Dosi, and L. Orsenigo. 1988. Innovation, diversity and diffusion: A self-organisation model. Economic Journal 98:1032 - 1054. Snyder, F. 1990. Ideologies of competition: Two perspectives on the completion of the internal market. In New directions in European Community law, edited by F. Snyder. London: Croom Helm. Stenberg, L. 1987. Utvetlingsblock i förnyelse av svensk industri. Stockholm: Industridepartementet. Stigler, G. J. 1964. A theory of oligopoly. Journal of Political Economy 72 (1):44 - 61. Stoneman, P. 1987. The economic analysis of technology policy. Oxford: Oxford University Press. Teubal, M., T. Yinnon, and E. Zuscovitch. 1991. Networks and market creation. Research Policy 20:381 - 392. Tovias, S. 1990. The impact of liberalizing government procurement policies of individual EC countries on trade with non-members. Weltwirtschaftliches Archiv (722). Trepte, P. A. 1993. Public procurement in the EC. Great Britain: CCH Editions Ltd. Tushman, M. 1977. Special boundary roles in the innovation process. Administrative Science Quarterly 22:587 - 605. Tushman, M., and P. Anderson. 1986. Technological discontinuities and organizational environments. Administrative Science Quarterly 31:439 - 465. Tushman, M. L., and W. M. Moore, eds. 1982. Readings in the management of innovation. Marshfield, Massachusetts: Pitman. Utterback, J., and W. Abernathy. 1975. A dynamic model of process and product innovation. OMEGA 3 (6):639 - 656. Utterback, J. M. 1994. Mastering the dynamics of innovation: How companies can seize opportunities in the face of technological change. Boston, Massachusetts: Harvard Business School Press. Utterback, J.M., and F. F. Suarez. 1993. Innovation, competition and industry structure. Research Policy 22:1 - 21. Utterly, J., and B. Harper. 1993. The political economy of competitive tendering. In The political economy of privatization, edited by D. Clarke and H. Pitelis. London: MacMillan. Vickrey, W. 1961. Counterspeculation, auctions and competitive sealed tenders. Journal of Finance 16 (March):8 - 37. Vogel, E. 1980. Japan as no. 1. Tokyo: Tuttle. von Hippel, E. 1988. The sources of innovation. Oxford: Oxford University Press. 148
Webster, F., and Y. Wind. 1972. A general model of organizational buying behaviour. Journal of Marketing 36 (2):12 - 19. Westling, H. 1996. Cooperative procurement: Market acceptance for innovative energy- efficient technologies. Stockholm: NUTEK B. Williams, R., and R. Smellie. 1985. Public purchasing: An administrative Cinderella. Public Administration 63 (1):23 - 39. Williamson, O. E. 1975. Markets and hierarchies: Analysis and anti-trust implications. New York: Free Press. WS Atkins Management Consultants, and Associates. 1988. The "cost of non-Europe" in public sector procurement. Edited by Commission of the European Communities. Vol. 5, Research on the cost of non-Europe -- Basic findings. Luxembourg: Office for Official Publications of the European Communities. Zysman, J. 1977. Political strategies for industrial order. Berkely: University of California Press.