This is a brilliant set of papers. Unlike many such collections by diverse authors, the standard is uniformly high and even more surprisingly, there is a clear common theme which links them together despite the lack of common authorship. That theme is the elucidation of just what is meant by ‘globalisation’. As Archibugi and Iammarino observe, this is a catch-all concept which is used indiscriminately to describe many diverse phenomena.

In particular, the book concentrates on the ways in which globalisation affects and is affected by technical change and systems of innovation. Over the last decade or somany authors have used the expression ‘national system of innovation’ to describe and analyse those networks of institutions and activities, which in any country, initiate, modify, import and diffuse new technologies. Some of the authors have attributed the origin of this concept to me. This is not accurate. To the best of my knowledge the expression was coined by Lundvall, who contributes the first chapter in this book in which he argues cogently that what matters most is learning, rather than knowledge itself. In any case, as I am sure he would agree, and as several of the chapters point out (e.g. Dosi and Kluth and Andersen) there is a long tradition in economic thought of this combined approach to technical innovation and institutional change, going back at least to Count Serra in Naples.

As this discussion has unfolded, it has become apparent that both the international (‘global’) and the sub-national (‘regional’) dimensions of innovative activities merit investigation and debate as well as the national dimension.

Introduction

We shall show that the activities generating the skills and know-how that give firms competitive advantage are less internationalised than all other dimensions of corporate activity. Even very large corporations in most cases perform most of their R&D at home. As a consequence, companies' innovative activities are significantly influenced by their home country's national system of innovation: the quality of basic research, workforce skills, systems of corporate governance, the degree of competitive rivalry, and local inducement mechanisms, such as abundant raw materials, the price of labour and energy, and persistent patterns of private investment or public procurement.

We do not foresee any fundamental changes in future. The efficiency gains from the geographical concentration of innovative activities will remain, even if firms increasingly seek out unique skills in foreign countries. The State will continue to provide infrastructure, incentives and institutions that strongly influence the rate and direction of innovative activities in locally based corporations. There are no mechanisms to ensure that national systems of innovation converge in either their characteristics or their performance. Nor are there any signs of the emergence of a European system of innovation: the dominant system is likely to remain the German one.

Other authors have come to similar conclusions, most notably Porter (1990), based mainly on in-depth case studies, and Hu (1992) based on an analysis of the nature and the behaviour of multinational firms. Our research is based on systematic information on the world's largest firms that were technologically active in patenting in the USA in the 1980s. It consists of 587 firms, of which 249 were US owned, 17 from Canada, 143 from Japan and 178 from Europe (Patel and Pavitt, 1991).

Introduction

National environments differ in their capability of stimulating, facilitating or preventing innovative activities of firms, and historically national economies and their institutional set-up have had a considerable influence on the firms' competitive success. A vast recent literature by formulating the concept of the ‘national system of innovation’ has confirmed that the structural characteristics of a national economy, such as its specific production structure, its technical infrastructure and other institutional factors, can strongly influence firms' innovative performances (Freeman and Soete, 1997; Lundvall ed., 1992; Nelson ed., 1993; Edquist ed., 1997). It has been shown that differences in national systems are particularly important between the United States, Japan and the EU, and between the European countries themselves.

The concept of national specificities determining national performance, however, has been recently challenged on the grounds that the current wave of globalisation and sophistication of financial markets are aligning national economies. The growing role of transnational corporations (TNCs) in the current global competitive phase, as many point out, is changing the face of the world economy in the direction of standardisation and convergence of national structures and performances. A few even predict that the nation state will soon be obsolete and that national diversities, that were very important in the past, are likely to disappear in the near future. The limiting case would be of a fully transnational economy without any residual disparities across countries (Ohmae, 1990).

This chapter assesses the issue of the linkage between globalisation and national specificities by looking at the dynamics of world trade and trade specialisation of major economies. It is divided into four sections.

All the chapters in this book were commissioned specifically for this volume and draft versions were discussed at a working conference in April 1996 in Rome. This conference marked the second in a series of Euroconferences entitled ‘The Globalization of Technology: Lessons for the Public and Business Sectors’ co-funded by DGXII of the European Commission as part of the Human Capital Mobility (HCM) Programme (Grant no. ERBCHECCT940230), organised by the three of us.

The overall Euroconference initiative has a number of objectives, but a key aim is to help inform, involve and support young scientists and researchers in the field of industrial innovation and technology policy. The Rome conference therefore sought to bring together an informal group of some young and some not so young researchers working in this field. The result was a lively and interesting debate surrounding the issues of national innovation systems and of the globalisation of technology which is of such crucial strategic importance to both private and public sectors alike.

Obviously a vital role was played by all the conference presenters, many of whom have subsequently become contributors to this book. We would therefore like to thank Jørn Andersen, Giovanni Dosi, Paolo Guerrieri, Simona Iammarino, Michael Kitson, Michael Kluth, Bengt-Åke Lundvall, Keith Pavitt and Mario Pianta for presenting papers and participating in the discussions. We would also like to thank John Cantwell, John Dunning, Pari Patel and Clifford Wymbs for their contributions.

We are grateful to the Institute for Studies on Scientific Research and Documentation of the Italian National Research Council for hosting the Conference in Rome.

The changing nature of foreign direct investment

Until about a decade ago, the received theory of foreign direct investment (FDI) asserted that firms established foreign value-added activities in order to exploit their home-based competitive (or ownership-specific (O)) advantages, and to benefit from the internalisation of cross-border intermediate product markets. To be sure, there was some natural resource seeking foreign direct investment (FDI), designed to gain access to agricultural products, raw materials and minerals not available, or not available on such advantageous terms, in the investing country. Multinational enterprises (MNEs) also invested in developing countries to take advantage of low cost and/or more productive unskilled or semi-skilled labour. However, in each of these cases, as with market and efficiency seeking FDI, the main objective of the investing firms was to seek out the location specific assets of foreign countries, which could be used in conjunction with their own mobile O specific assets, which, in turn, were assumed to reflect the availability and quality of created assets in their home countries.

World economic events of the last decade – particularly the globalisation and liberalisation of markets, a new generation of technological and organisational advances and the emergence of third world countries as significant sources of FDI – have changed the economic environment for MNE activity. While, as with exports and licensing arrangements, international production continues to be a means of capturing economic rent on the O advantages of the investing companies, increasingly firms are investing abroad to protect or augment their core competencies.

Introduction

In these short notes it is certainly impossible to provide any fair assessment of the wealth of research which has gone over the last decade into the analysis of national systems of innovation (and, relatedly, national systems of production). Hence, in the following, I will limit myself to some rather telegraphic remarks and propositions concerning, first, the empirical background of such studies; second, the relations among (partly different) interpretative perspectives on the subject; and, third, their implications also for other domains of economic analysis – including the theory of the firm and growth theory. Finally, I shall briefly hint at some policy implications, especially with reference to the European case. In doing all that I will raise more questions than provide answers: however, a few of the propositions that follow are empirically testable, and thus may provide some inspiration for further research in the area.

Persistent asymmetries across firms and countries

The general historical background of the discussion of ‘national systems’, as I see it, is the observation of non-random distributions across countries of:

1. corporate capabilities;

2. organisational forms;

3. strategies; and ultimately

4. revealed performances, in terms of production efficiency and inputs productivities, rates of innovation (however measured), rates of adoption/diffusion of innovation themselves, dynamics of market shares on the world markets, growth of income and employment.

Note that the patterns defined by the latter indicators, when measured at the level of sectors and countries, tend to display relatively high degrees of persistence over time, despite a somewhat higher inter-company variability.

Introduction

This chapter explores the possible existence and form of regional systems of innovation (RSI). Christopher Freeman has defined a national system of innovation (NSI) as the network of institutions in the public and private sectors whose activities and interactions initiate, import, modify and diffuse new technologies (Freeman, 1987, p. 1). Although this definition was applied at a national level it can arguably be equally applied at a regional or local level. This chapter will seek to explore these issues, not suggesting that regional systems of innovation should be seen in some way as supplanting national systems of innovation, but rather should be viewed as providing another layer or conceptual lens to the whole system of innovation. In so doing, it seeks to develop in a geographical sense at least part of Metcalfe's (1995, p. 41) view that the national unit may be too broad a category to allow a clear understanding of the complete dynamics of a technological system and instead focus should be on ‘a number of distinct technology-based systems each of which is geographically and institutionally localised within the nation but with links into the supporting national and international system’.

More specifically, therefore, this analysis will examine whether the broad definition of national systems of innovation can also be applied at a regional level. Do, or can, regions offer distinct systems of innovation that are worthy of study? If they do exist, in what ways are they different from national systems of innovation? If the concept is relevant, is it becoming more or less applicable over time? What more general lessons can be learnt from the study of regional systems of innovation?

The introductory discussion provides an example to illustrate the possibility of coordination failure due to the presence of multiple Pareto-ranked Nash equilibria. Of course, as remarked earlier, whether coordination failures actually occur depends on the selection of an outcome from the set of Nash equilibria. If, for instance, the Pareto-dominant Nash equilibrium is a natural focal point, as suggested, for example, by Harsanyi–Selten [1988], then understanding the macroeconomic implications of coordination failures would be somewhat less interesting.

This chapter first reviews some recent experimental evidence that bears directly on equilibrium selection in coordination games. The evidence both concerns results on equilibrium selection and provides some insights into the process of equilibration. The second part of this chapter describes a variety of selection theories that bear directly on coordination games.

One conclusion of the experimental evidence is that coordination problems are not a pure theoretical curiosity. In particular, coordination failures are routinely observed in experimental games.

EXPERIMENTAL EVIDENCE

The discussion of experimental evidence is partitioned into three parts. First, evidence from simple coordination games is presented. Given the frequency of coordination failures observed in these experiments, further treatments which explore possible remedies to coordination problems, such as preplay communication, are presented.

Baseline Experiments

As a starting point, consider the coordination game in Figure 1.1. This game was the focus of a study of experimental coordination games by Cooper, DeJong, Forsythe and Ross [1992]. In this game, there are two pure strategy Nash equilibria, {1, 1} and {2,2}, and a mixed strategy equilibrium.

The next class of models stresses interactions between agents that reflect various forms of trading externalities. Here the linkages between agents do not arise directly in preferences or the production process but rather in the way agents come together to trade. Thus, these models rest firmly on the view that the Walrasian auctioneer does not function to bring together traders. Instead, trade frictions arise from the process of search and recruitment.

The complementarity in these models stems from the “thick market.” This is essentially a restriction on the relationship between trading costs and the level of economic activity. In particular, the economies we explore have the property that if there are many agents in the market searching for trading partners, then the returns to participating in the market are higher as a result of reduced costs of search. Thus thick market effects are just the opposite of congestion effects: the thicker the market the lower are trading costs.

AN EXAMPLE

The flavor of these models can be seen through a simple example of a participation complementarity. Denote by Z(p) the return to an individual from participating in an activity if a proportion, p, of others are participating. Assume that Z'(·) > 0 so that thicker markets are more desirable. The economics underlying this assumption will be the focus of the discussion to follow. Suppose that agent i's opportunity cost of participating is ki and that these costs are distributed across the population with a cumulative distribution function given by H(k), for k ε [0, 1].

The goal of the previous chapter was to provide an introduction to coordination games and some evidence on coordination failures. In the end, we find that coordination failures can arise in fairly simple experimental settings. While these experiments are certainly suggestive that coordination problems may arise, they leave open an important question: what are the underlying economic interactions that lead to coordination games?

The answer offered in this chapter takes the form of two abstract frameworks for analysis. The first, drawing upon Cooper and John [1988], stresses the interaction between agents in strategic settings where strategies are simply scalars in a closed interval. This formulation leads to a relatively straightforward equilibrium analysis, including conditions for multiple equilibria and some welfare results. The main point is that coordination games, such as those illustrated in the previous chapter, rest upon an interaction between agents termed strategic complementarity. As suggested already, this interaction implies that increased effort by other agents leads the remaining agent to follow suit. Besides becoming the basis for multiple equilibria, the strategic complementarity gives rise to multiplier effects.

The second part of the chapter looks at more general interactions. While almost all current macroeconomic applications of coordination games can be cast in the Cooper–John framework, the more general structure, investigated most recently by Milgrom and Roberts [1990] and Vives [1990], is quite powerful and worthy of study.

COOPER–JOHN MODEL

Cooper and John [1988] consider a game which highlights the key theme in this literature: the concept of strategic complementarity.

In this final section of the book we are interested in studying the problem of policy determination between a government and a set of private agents. The inefficiencies created by the presence of external effects as well as the prospects of multiple equilibria studied in the previous chapter set the stage for a consideration of government intervention to resolve these problems. Thus this topic is a natural conclusion to our study of macroeconomic complementarities.

The starting point of the chapter is an illustration of the coordinating power of the government. If coordination problems reflect the inability of agents to select the Pareto-optimal (optimistic) Nash equilibrium, then the government may be able to take actions to achieve the desired outcome. As we shall see, the government's actions can eliminate some undesirable equilibria by turning the strategies that support them into dominated strategies. These policies can be thought of as “confidence building measures” that work by eliminating the pessimistic beliefs that support the Pareto-inferior (pessimistic) Nash equilibria.

One important theme here is that in the optimistic equilibrium, the government never takes an action. Instead, its commitment to an action is sufficient for stabilization through removal of the pessimistic equilibrium. Thus governments may appear to be doing “nothing” when, in fact, they are quite successful.

To illustrate, we study the Diamond and Dybvig [1983] model of bank runs and the role of the government in supporting the Pareto optimal equilibrium through the creation of deposit insurance.

We begin the study of economic environments underlying coordination games by considering the most direct form of interaction across agents: through a production function. As we shall see, this simple structure forms the basis for new insights into both aggregate economic fluctuations and growth. Further, this source of complementarity is most tractable in terms of quantitative analysis since it is most easily incorporated into the stochastic growth model.

Consequently, the discussion in this chapter contains both theory and quantitative evidence associated with the behavior of these economies. This focus reflects, in fact, recent developments in quantitative analysis which allow us to go beyond the stochastic growth model studied by Kydland and Prescott [1982] and King, Plosser and Rebelo [1988] to understand macroeconomic dynamics of economies with distortions.

INPUT GAMES AND TECHNOLOGICAL COMPLEMENTARITY

Assume that I agents provide effort into a joint production process. The per capita output of this process is f(e1, e2, …, eI) where ei is the effort level of agent i in the production process. We assume that e ∈ [0, 1] so that the strategy space is a complete lattice. Per capita output is also the consumption for each agent. Implicit here is an assumption about the nature of the compensation scheme: agents share equally in the output from their joint production.

Let U(Ci) be the utility from goods consumption and g(ei) be the disutility of effort for agent i. Hence

σ(ei, e) = U(f(e1, e2, …, eI)) – g(ei) (1)

Assume that U(-) is strictly increasing and concave and that g(-) is strictly increasing and strictly convex.

The goal of this book is to provide a synthesis of research on the topic of complementarities in macroeconomics. Its primary goal is to isolate the various sources of complementarity and then to explore their implications for the behavior of macroeconomies. The success of this approach is seen through the numerous theoretical and empirical applications of the basic structure inherent in model economies built upon the macroeconomic complementarities structure.

As this is principally a book about applications in macroeconomics, it has been necessary to leave aside a number of topics that relate to the implications of complementarities for other branches of economics, such as industrial organization. Still, the reader interested in applications outside macroeconomics ideally will find the more general discussion of models of complementarities as well as the presentation of experimental evidence of some value.

The first two chapters as well as the next section of this Preface focus on general issues arising in models of complementarities, thus providing a framework for the more applied analysis that will follow. In particular, the first two chapters discuss experimental evidence on coordination games and theories of selection and put forth a general model of macroeconomic complementarities.

The remaining chapters explore applications of the general structure by investigating particular channels of interactions across agents. This includes the study of economies in which (i) externalities are present in the technology of the individual agent, (ii) markets are imperfectly competitive, (iii) agents come together through a search process and (iv) information is imperfect. As we shall see, all of these deviations from the standard general equilibrium model of Arrow and Debreu can give rise to macroeconomic complementarities.

One feature of aggregate behavior is the synchronization of discrete decisions such as the purchase of durable capital by firms and durable goods by households. These expenditures are important to understand in that they are extremely volatile elements of aggregate spending. Put differently, these are the elements of total expenditure that display the most time series variance. Introducing these discrete choices into traditional general equilibrium models is somewhat difficult because of the nonconvexity associated with lumpy expenditures on consumer and firm durables. One means of dealing with these nonconvexities is to look for equilibria in which there is some “smoothing by aggregation.” The effect of doing so, however, is that these discrete choices are, by construction, no longer synchronized so that their macroeconomic importance is dramatically reduced. In fact, this approach works only if agents have an incentive to take actions at different points in time.

When, in contrast, agents have an incentive to synchronize their discrete choices, then smoothing by aggregation is no longer possible. In this case, synchronized discrete decisions can matter for the macroeconomy. They can create endogenous fluctuations of the aggregate economy and magnify underlying disturbances to that economy.

The focus of this chapter is on the basis for synchronization and, more generally, the issue of the timing of economic activity. The first part of the chapter looks at incentives for the synchronization of activity. The second part goes on to explore the issue of delay.