Introduction

Untill recently, the modern literature on geography and trade paid relatively little attention to the relationship between agglomerating and spreading forces on the one hand and the structure and volume of (international) trade on the other. International trade flows are undoubtedly largely determined by the spatial distribution of economic activity. Taking the core (symmetric) two-country model of chapters 3 and 4 as a point of departure, the predictions on the structure and size of trade flows are simple. If economic activity is evenly spread, food is not traded internationally, so there is only intra-industry trade of manufactures between the two countries. If there is complete agglomeration of manufacturing activity, the only other possible long-run outcome, there is exclusively inter-industry trade (food for manufactures) between the two countries. Although these basic predictions are in line with empirical observations, that is trade is large between similar countries and dominated by intra-industry trade (see Box 9.1), the basic structure is too extreme in its predictions and too rigid in structure to allow for different types of international trade flow. The objective of this chapter is to demonstrate how international trade models may be combined with the geographical economics structure to allow for a diversified and rich explanation of international interactions. In doing so it partially fills the gap in the literature observed by Bertil Ohlin in 1933, namely the need to develop a theory of location which may serve as a background for a theory of international trade; see chapter 1.

Introduction

Chapter 3 has carefully developed and discussed the main features of the core model of geographical economics. Most importantly, perhaps, the model is coherent: it is a miniature world in which the demand in one region for the manufactures of another region is not imposed beforehand but derived from the income generated in the region by its production and exports to other regions. Despite the care taken in setting up the different aspects of the model as simply and tractably as possible it turns out to be quite complex to study analytically.

This chapter focuses on the possibilities and the advantages and disadvantages of computer simulations to better understand the workings of the core model of geographical economics. In doing so, we will be able to say more on the determination of the long-run equilibrium in the core model. We start by explaining in some detail what computer simulations are, what they can and cannot do, and the specification issues involved in getting them to work. Subsequently, we show how computer simulations can be useful in three ways. First, we demonstrate that simulations allow us to do things one cannot do analytically. In general this is indispensable in more complicated models to get a “feel” for the model. The most important goal of these simulations is to see how certain crucial aspects of the model react to changes in important parameters of the core model. Second, analytical solutions are sometimes possible and simulations can give rise to ideas which can be proven analytically. Of the latter we give two examples, in sections 4.6 and 4.7 below. Third, as we show at the end of this chapter, simulations can be useful in demonstrating that certain ideas or suggestions do not always hold, simply by producing a counter-example.

Introduction

As has already been mentioned in the preface, Bertil Ohlin (1933) observed long ago that the fields of trade theory on the one hand and regional and urban economics on the other had in principle the same research objectives. Both areas want to answers the questions: who produces what, where, and why. Despite Ohlin's observation, each field has continued to go its own way since the nineteenth century. Chapter 2 showed that trade theory assumes that countries are dimensionless points in space. Trade theorists are mostly interested in how market structure, production techniques, and consumer behavior interact. The resulting factor and commodity prices determine the pattern of international trade flows. Location is at best an exogenous factor and usually does not play a role of any significance. Regional and urban economics, instead, take market structure and prices as given and try to find out which allocation of space is most efficient. The underlying behavior of consumers and producers, central in trade theory, is less important (Fujita and Thisse, 1996). Although both strands of literature produce valuable insights in their own right, trade theory and regional and urban economics are productively combined in geographical economics.

This chapter discusses and explains the core model of geographical economics, a small general equilibrium model developed by Krugman (1991). As we shall see, the equilibrium equations of this model are non-linear. This means that small changes in parameters do not always produce the same effects; sometimes the effects are small, sometimes they are large.

Introduction

The core model of geographical economics was introduced and explained in chapters 3 and 4 of this book. In our analysis of various extensions and applications in chapters 5–10 we investigated relatively small modifications of the core model, usually only affecting the cost function, and thus the production structure, of the core model. This was done on purpose. Not only for didactic reasons (each time returning to the familiar territory of the core model),but also to demonstrate that (seemingly) small changes in the core model can drastically increase its applicability and have interesting and sometimes far-reaching consequences. In our discussions of these adaptations of the core model two important questions have, however, been unduly neglected. First, what are the policy implications (if any) that arise from the core model and its extensions? Or, more generally, can geographical economics be used for policy analysis? Second, now that we have come to the end of our inquiry into geographical economics, what are the strong and weak points of this approach? In other words, how should geographical economics be assessed? The closing chapter of this book therefore deals with the policy implications and with the value-added of geographical economics. The latter is, inevitably, a subjective undertaking, but it gives us the opportunity to express our own views on the advantages and disadvantages of geographical economics.

This chapter is organized as follows. In the next section, we will conduct a simple policy experiment by doing simulations with the congestion model of chapter 7 in a world of non-neutral space.

Introduction

It happened on October 12, 1999, at least according to the United Nations (UN). From that day more than 6 billion people have inhabited the planet Earth. Of course, given the inaccuracy of the data, the UN could have been off by 100 million people or so. Every day some 100 million billion sperms are released and 400,000 babies are born, whereas “only” 140,000 persons die. Consequently, the world population has been growing rapidly, especially over the second half of the twentieth century. Given the average population density in the world, about 44 persons per square kilometer, if you are part of a family with two children, your family could have about 9 hectares (or 22.5 acres) at its disposal. The large majority of our readers will probably look around in amazement to conclude that they do not own an area close to this size. The reason is simple: the world population is unevenly distributed. But why?

There may be many reasons why people cluster together: sociological – you like to interact with other human beings; psychological – you are afraid to be alone; historical – your grandfather already lived where you live now; cultural – the atmosphere here is unlike anywhere else in the world; geographical – the scenery is breathtaking and the beach is wonderful. At best we will discuss the above reasons for clustering only cursorily, because our attention in this book will be focused on the economic rationale behind clustering or agglomeration.

Introduction

The central message of chapter 1 is that geography is important. Economic activity is not evenly distributed across space. On the contrary, clustering of economic activities can be found at various levels of aggregation: the considerable variation in economic size of cities or regions at the national level, or the uneven distribution of wealth and production at the global level. The question arises, of course, of why location apparently is relevant for the distribution of economic activity. To answer this question, we need an analytical framework in which geography plays a part one way or another. In particular, we would like to show that the decisions of economic agents are determined by geography, and that geography itself can be derived from the behavior of economic agents. This is, in a nutshell, what the approach developed in this book tries to do. We want to make absolutely clear from the start that this approach, referred to throughout this book as geographical economics but (see the preface) perhaps better known as new economic geography, is by no means the first theory to address location issues. There is a long tradition that deals with these questions, and this will be discussed in this chapter. The novelty of geographical economics is not the research topic, but the way it tackles the relationships between economics and geography.

Before we turn to the core model of geographical economics, this chapter discusses the role of geography in economic theory. It is clearly beyond the scope of this book to try to give a complete survey of the literature. Instead, we highlight the role of geography in some important fields of economics.

Introduction

So far we have not paid much attention in this book to the intermediate dynamics underlying the geographical economics models. Instead, we have usually focused attention on the relationship between a long-run equilibrium outcome and the structural parameters, given an initial geographical distribution of labour and production. For that reason we argued in chapter 2 that the novelty of geographical economics relative to new trade theory is to be found in the endogenous determination of market size, fostered by the migration of mobile workers towards regions with higher real wages. The dynamics underlying the adjustment path, that is how we evolve over time (see our remark below on “time”) from an initial distribution to a final distribution, and the intricacies of economic growth and development have been virtually absent in the analysis so far. This chapter partially fills this void. In doing so we will distinguish among three types of dynamics, increasing in complexity and in importance:

1. (i) adjustment dynamics

2. (ii) simulation dynamics

3. (iii) economic growth.

Adjustment dynamics

This type of dynamics analyzes the adjustment path over time, from an initial distribution of manufacturing production across regions to a final long-run equilibrium, by showing the sequence of short-run equilibria leading to the long-run equilibrium.

Introduction

Globalization has many faces. The most characteristic aspect of globalization is that it appears that the world becomes smaller; transport costs are reduced, trade barriers disappear, and information becomes less expensive and becomes itself an internationally traded product. Although spreading of economic activity is certainly possible, the lessons of the geographical economics approach might increase fears of an ever-growing gap between nations, in which, because of decreases in trade costs, center–periphery structures become the rule instead of the exception. One of the major actors in this process of globalization is the multinational firm, or multinational for short. This is probably the most mobile among all firms, with sufficient “international” knowledge to seize a profitable opportunity when it presents itself. Without specific cultural ties to individual nations, it can rapidly move in and out of countries, acting only on economic incentives. Given the recent growth in foreign direct investment (FDI), which at the moment is growing faster than international trade, one might expect multinational firms to be decisive in at least some of the agglomeration and spreading trends going on today. It will turn out that the geographical economics approach is an excellent and promising methodology to look at multinationals. Together with the general equilibrium models developed in the 1980s by Helpman and Krugman (1985), they provide an improvement to the useful taxonomic OLI framework of Dunning (1977, 1981); see section 8.3.

In chapter 6 we explained how the introduction of intermediate goods can change the mechanisms leading to spreading or agglomeration. Firms, rather than labor, were assumed to be mobile. Agglomeration can come about if workers move to other sectors; in other words, workers are immobile between regions or countries, but mobile within those regions and countries.

Introduction

In the first chapter of this book a number of stylized facts about the clustering of economic activity were presented to justify our inquiry into the relationship between economics and geography. Chapters 2–4 have mainly looked at this relationship from a theoretical point of view. In this chapter we start with the reminder that in reality there is a considerable degree of clustering, that is location matters. Our main objective in this chapter is to assess the empirical relevance of geographical economics and to see whether and how the theoretical model of chapters 3 and 4 can be tested.

In the next section we briefly review the main facts about the concentration and agglomeration of economic activity at different levels of aggregation. This continues, and partly restates, our discussion of stylized facts of location in chapter 1. Against this background, section 5.3 deals with the question of whether these facts can be reconciled with the various economic theories of location presented in Chapters 2–4. To answer this question, we summarize the main predictions about clustering (if any) that follow from the various theories, and conclude that the stylized facts are in accordance with several theories (and not only with geographical economics). This does not come as a big surprise since many empirical studies about the concentration or the agglomeration of economic activity are simply not primarily concerned with the testing of specific theories.

In sections 5.4 and 5.5 we analyze two recent attempts which explicitly try to test (part of) the implications of the geographical economics models. The bulk of this chapter consists of an in-depth discussion of these two attempts, such that we will not offer a fully-fledged survey of the empirical aspects of trade, growth, and location.

Introduction

The purpose of this chapter is not to provide an exhaustive account of the various theories of the state found throughout the already burgeoing constructivist literature. Given that new variants are only now emerging, my purpose must be more modest. Here I present three distinct theories of the state, which cover many, though not all, of the variants found in the extant constructivist literature. Unlike previous chapters, I shall begin with a general introduction to constructivism, before proceeding to examine three key variants and their approaches to the state.

Constructivism versus ‘rationalism’

In the past, IR scholars have tended to divide IR theory into three competing schools: liberalism/pluralism, realism and Marxist structuralism (Banks 1985). But the rise of constructivism reconfigures the traditional ‘trichotomy’ into a dichotomy, comprising constructivism on the one hand and ‘rationalism’ or ‘neo-utilitarianism’ on the other. Now realism, liberalism and Marxism are all placed in the one category of ‘rationalist’ or ‘neo-utilitarian’ approaches. Constructivists critique ‘rationalist’ theory on many grounds.

Constructivists begin by arguing that rationalist theory (not to be confused with English school rationalism) has been excessively materialist and agent-centric. For rationalists, IR appears as the product of agents (usually states) which are imbued with ‘instrumental rationality’.

Introduction: the two realisms of international relations theory

Conventional wisdom conflates neorealism and classical realism (e.g. Gilpin 1986; Grieco 1993a: 135). But my interpretation suggests two clearly differentiated realisms and two distinct theories of the state, as revealed within the framework of the second state debate. These two positions are juxtaposed in figures 2.1 and 2.7 (p. 46). There is a relatively strong consensus among realists and non-realists as to what constitutes ‘neorealism’. I summarise the approach through ‘six principles’, outlined on the left-hand side of figure 2.1. In essence, neorealism is highly parsimonious, such that although the state has high domestic agential power (or high institutional autonomy), nevertheless it has no international agential power to determine policy or shape the international system free of international structural constraints. For neorealism, states are in effect ‘passive bearers’ (Träger) of the international political structure. This contrasts with what I call ‘the six principles’ of classical realism which boil down to the essential claim that while states' domestic agential power varies through historical epochs, nevertheless all states have at all times (albeit to varying degrees) sufficient levels of international agential power to shape the inter-state system. Both Carr and Morgenthau emphasise that, under certain circumstances, states can create a peaceful world. Morgenthau (1948/1978: chapter 32) argues that the regaining of high domestic agential power can enable the state to create the necessary conditions for a peaceful world (i.e. enable the generation of high international agential state power).

Introduction: the ‘two waves’ of neo-Weberian historical sociology

In the last decade, IR as a discipline has been undergoing a crisis as its master-paradigm, neorealism, is increasingly seen as limited, if not obsolete. Accordingly, IR is thought to have reached an impasse (Ferguson and Mansbach 1988; Halliday 1994). Why? It is perceived by critics that neorealism has many blind spots, four of which are: a lack of a theory of the state and an exaggeration of ‘structure’ to the detriment of ‘agency’ an inability to theorise the integrated nature of global politics, given the assumption that there is a fundamental separation or dichotomy between the international and national realms; a lack of a theory of international change; and a static a-historical approach. One response to the crisis of neorealism made by some IR scholars has been to turn towards neo-Weberian historical sociology (WHS) to provide a way out of the impasse (e.g. Jarvis 1989; Halliday 1994; Hobson 1997, 1998a; Hall 1998; Hobden 1998; Seabrooke 2000). Thus WHS is thought to offer a theory of the state that is allegedly missing in neorealism, given that WHS problematises the state, and seeks to describe and explain its origins, powers and changing configurations over time. This also allegedly offers a means to go beyond neorealist structuralism by bringing ‘agency’ back in. Secondly, WHS advocates an intimate relationship between the internal and external realms, thereby offering a potentially rich ‘integrationist’ approach. While Waltz actually recognises that the international can shape the national, he nevertheless dismisses the possibility that the national can shape the international realm. Thirdly, WHS provides a theory of change, which allegedly helps counter neorealism's static approach.