Handbook of Regional Science

This chapter traces the development of the role of economic opportunities in the study of migration. From the earliest years of internal migration as a recognized field of study, scholars in many social science disciplines believed that such opportunities were key determinants of migration. However, during the late nineteenth and early twentieth centuries, the lack of statistical measures of income and wages at subnational levels prevented empirical testing of the economic opportunity hypothesis. During this time, much rural-to-urban migration was occurring, and the presumption was that these flows were being driven by perceived urban–rural differences in economic well-being. The first formal measures used by economists in the 1930s were regional unemployment rates, and these rates proved to be significant determinants of migration during the Depression, but did not always hold up to scrutiny in later years. As aggregate income measures became increasingly available after 1960, they were incorporated in migration models, but their empirical success also was limited. Finally, the availability of microdata that reflects personal employment status and household income has allowed numerous advances in our understanding of various migration phenomena and also has helped clear up many dilemmas regarding earlier migration studies that used aggregate data.

Over long periods of human history, labor market equilibrium involved movements from low-wage areas to high-wage areas, a form of arbitrage under the implicit view that wage differentials corresponded to utility differentials. This “labor economics” view is likely to be viable as long as movement and information costs are high, and under this view, the movements would be expected to cause wage convergence over space. In recent decades, beginning as early as the 1960s in the United States, both the out-of-pocket and psychological costs of movement have plummeted with advances in transportation and communication technology and innovation. In addition, these same advances have enabled individual households and firms to have vastly improved information about potential benefits of locating in a host of potential locations. These observations, along with recent failures to observe convergence in wage rates, suggest that an alternative view – assuming a utility equilibrium over space – might better predict and explain the labor market equilibrium. This “urban/regional economics” view takes wages and rents as being compensatory for varying levels of household and firm amenities. In this view, whether the spatial equilibrium in labor markets involves convergence or divergence becomes quite a complicated issue. This chapter explores a number of the complexities, hinting at a broad range of potentially fruitful future research.

This chapter reviews labor supply, demand, and equilibrium topics with the goal of showing how they determine labor market area (LMA) outcomes across geographic space. Labor supply curves are based on utility-maximizing choices between working and leisure, subject to a budget constraint, while labor demand curves are derived from the firm’s production function assuming profit-maximizing behavior. The challenges of defining and empirically delimiting LMAs are examined from historical perspectives and using statistical clustering analysis, with commuting data serving as a key tool. A key distinction is drawn between functional versus homogenous regionalization problems, and a number of suitable statistical approaches are reviewed. Current models used to study differences in earnings across labor markets as well as the effects of boom and bust cycles are also discussed. An empirical technique is presented for decomposing employment change within a community into four key labor market concepts: commuting, unemployment, labor force participation, and migration.

This chapter summarizes the main developments in job search theory ever since its inception in the 1970s. After describing the assumptions and formulation of the basic model, the chapter moves onto analyzing how the original framework has been extended by removing some of the initial limitations. A separate section is then devoted to the matching function theory which represents one of the main developments of job search theory in more recent years and whose importance has been recognized by the award of the 2010 Nobel Prize in economics. The last section attempts to reconcile job search and migration theory by introducing the role of space and describing the main contributions on these topics by regional economists.

In the monocentric model, commuting is viewed as a burden whose cost shapes the spatial structure of cities to a considerable extent. This view has been challenged by the finding that actual commuting patterns are far from efficient. However, this “wasteful” commuting is better interpreted as an indication of labor market frictions that are traded off against commuting frictions than as a neglect of commuting costs. Urban sprawl results from the decreasing importance of physical space that was the consequence of the automobile and is fundamentally consistent with the basic insights of the monocentric model. Large and diversified urban labor markets flourish when space restrictions are relaxed because this facilitates the matching of jobs and workers along other dimensions. Having a large mortgage puts more stress on this allocation mechanism.

Spatial mismatch relates the unemployment and poverty of vulnerable population groups to their remoteness from job opportunities. Although the intuition initially applied to African Americans in US inner cities, spatial mismatch has a broader validity beyond the sole US context. In light of a detailed presentation of the mechanisms at work, we present the main results from various empirical tests of the spatial mismatch theory. Since key aspects of that theory remain to be tested, we also discuss methodological approaches and provide guidance for further research. We derive lessons for policy implications and comment on the appropriateness of related urban policies.

A prominent theme in the socioeconomic and regional science literature has been the topic of unemployment. We focus on regional unemployment and put forward candidate series of explanations for it using a basic model of labor supply and demand. The persistence of regional unemployment differentials points to inefficiencies in labor markets that in the long run could affect aggregate unemployment rates. Both a lack of labor demand and a constraint of labor supply increase regional unemployment. We finally discuss people- and place-based policies which aim to reduce high unemployment rates.

This chapter presents a comprehensive review of the fundamental concepts regarding real estate and housing markets. It aims firstly to provide an overview of the specific features of real property in general and housing in particular that make property a unique and multidimensional “good.” Building upon that, the chapter presents the key analytical tools extensively used in the relevant literature to capture the functioning of the real estate market as a set of interconnected markets, namely, the user (or space) market, the capital (or investment) market, and the development market. In this context, property development is examined as a process serving to reconcile long-run demand and supply imbalances generated in the user and investor markets. With regard to the housing market, after an overview of the key determinants of housing demand and supply, this chapter places its focus on the link between housing and the macroeconomy. Finally, the chapter explores the role of financial internationalization in the operation of real property markets and housing in particular, in the context of an increasingly globalized economy.

This chapter explores the literature on residential mobility and house price hedonics. Residential mobility studies the decision of economic agents to move or not and, if they move, their choice of new residence. Topics covered in this chapter include the theory behind the move-or-stay decision, modeling intra- and interregional moves, empirically validated determinants of moving, and macro- and microlevel studies on mobility. Next, house price hedonics explain the price of a house as the sum of all the things that give a house value, from structural characteristics like the number of full bathrooms to public services and neighborhood characteristics that the house experiences. The chapter discusses the theory behind hedonics, applications of the technique, and empirical approaches to identify hedonic house price studies and second-stage hedonic regressions of the demand and supply of characteristics that give a house its value.

This chapter provides an overview of the literature on neoclassical growth, starting with the simple Solow-Swan model and highlighting the main components of the neoclassical growth process. It considers the assumptions, predictions, and limitations of the Solow-Swan model and discusses several extensions that address some of these limitations and, in particular, those that are unrealistic for a regional growth setting. Several more complex models are presented and discussed, including a model that allows for exogenous technological progress, one that includes a broader definition of capital to also encompass human capital, and one that relaxes the assumption of a closed economy. Finally, the chapter considers a more complex model of neoclassical growth, the Ramsey-Cass-Koopmans model, which incorporates consumer behavior and allows for an endogenously determined savings rate.

In this chapter, we outline the basic mechanisms in endogenous growth theory that identify knowledge creation and diffusion as the core driver of economic growth. Then we discuss how new economic geography, urban economics, organizational science, and entrepreneurship theory have regionalized the mechanisms involved. Knowledge creation, however, has been dubbed a proximate cause of growth, and the quest for fundamental causes has continued. We then discuss this recent development in macroeconomic growth theory and argue that the new institutional approach to growth opens up a lot of new avenues for further research. Once again, the importance of cities, organizations, and entrepreneurship is being ignored in macro growth theory. Yet economic geography, urban economics, organizational science, and entrepreneurship theory have a lot to contribute to growth theory by both empirically and theoretically developing our understanding of local institutions and linking these to regional economic development and growth.

We describe how endogenous growth theory has now incorporated spatial factors. We also derive some of the policy implications of this new theory for growth and economic integration. We start by reviewing the product variety model of endogenous growth and discuss similarities with modeling techniques in the new economic geography. Both use Dixit-Stiglitz competition. Increasing returns provide an incentive for innovation in endogenous growth theory, and in combination with transport costs, increasing returns provide an incentive for firm location decisions in the new economic geography. Since innovation is the engine of growth in endogenous growth models and knowledge spillovers are a key input to innovation production, we also explore how innovation and knowledge have distinctly spatial characteristics. These modeling similarities and the spatial nature of knowledge spillovers have led to space being incorporated into the theory of endogenous growth. We guide the reader through how space is modeled in endogenous growth theory via the new economic geography. Growth by innovation is a force for agglomeration. When space is included, growth is enhanced by agglomeration because of the presence of localized technology spillovers. We consider the many other spatial factors included in models of space and growth. We explore the spatial effects on economic growth demonstrated by these theoretical models. Lastly, we consider policy implications for integration beyond lowering trade costs and discuss how lowering the cost of trading knowledge is a stabilizing force and is growth enhancing.

Oligopolies are a fundamental economic market structure in which the number of competing firms is sufficiently small so that the profit of each firm is dependent upon the interaction of the strategies of all firms. There are alternative behavioral assumptions one may employ in forming a model of spatial oligopoly. In this chapter, we study the classical oligopoly problem based on Cournot’s theory. The Cournot-Nash solution of oligopoly models assumes that firms choose their strategy simultaneously and each firm maximizes their utility function while assuming their competitor’s strategy is fixed. We begin this chapter with the basic definition of Nash equilibrium and the formulation of static spatial and network oligopoly models as variational inequality (VI) which can be solved by several numerical methods that exist in the literature. We then move on to dynamic oligopoly network models and show that the differential Nash game describing dynamic oligopolistic network competition may be articulated as a differential variational inequality (DVI) involving both control and state variables. Finite-dimensional time discretization is employed to approximate the model as a mathematical program which may be solved by the multi-start global optimization scheme found in the off-the-shelf software package GAMS when used in conjunction with the commercial solver MINOS. We also present a small-scale numerical example for a dynamic oligopolistic network.

In this chapter, we focus on theories and models of growth that have their origin in Keynesian economics. Their common features are that firstly, growth is largely export-driven; secondly, increasing returns yield path dependencies and possible divergence; thirdly, full resource utilization is not guaranteed; fourthly, economic expansion may face a balance of payments constraint, even at the regional level; and fifthly, institutions matter. We first briefly contrast demand-driven growth theories with neoclassical and other perspectives in taxonomy of growth theories. We then show how growth in exports yields regional income growth via a multiplier that is positively associated with the propensity to consume locally produced output and the propensity to invest but negatively related to regional tax rates and the extent to which government transfers are countercyclical. We show that Verdoorn’s law – economic expansion generates productivity growth – leads to both sustained export growth and steady-state income growth, with the latter in balance of payments equilibrium equaling the rate of growth of exports divided by the income elasticity of the demand for imports. Next, theories are reviewed that suggest that policies that encourage regional growth in wages and public expenditure can be growth enhancing. Finally, we argue that the effectiveness of such demand-driven growth policies depends on institutional settings.

Our understanding of people’s well-being was, until very recently, inferred from observable objective indicators such as their income and education. These measures were then aggregated to generate an average that characterized the city or region. With the growing availability of sample survey data, we now have at our disposal an increasing range of subjective measures of well-being that capture quality of life assessments made by individuals themselves. It is these internal measures of subjective well-being from microdata that are now being widely used throughout the social sciences to study what we call well-being or “happiness.”Contemporary interest in subjective measures of well-being stems from a wish to supplement market-based criteria such as GDP per capita with other more direct measures of societal well-being. Subjective measures are particularly useful in areas where the distribution of outcomes is not easily identified using other, especially market, criteria. The effect of investment in public infrastructure or the provision of green space or in fostering community networks or in redeveloping neighborhoods can be captured in responses to questions on well-being, preferably over time. These subjective measures, which have been shown to be highly correlated with clinical and other assessments of well-being, are likely to be of particular interest in regional science because of the way changes to places result from, or generate, a range of positive or negative externalities.

This chapter provides a selective survey of the main developments related to the study of regional convergence. We discuss the methodological issues at stake and show how a number of techniques applied in cross-country studies have been adapted to the study of regional convergence. In doing this, we focus on the two main strands of growth econometrics: the regression approach where predictions from formal neoclassical and other growth theories have been tested using cross-sectional and panel data and the distribution approach, which typically examines the entire distribution of output per capita across regions. In each case, we show how the analysis of regions rather than countries emphasizes the need to take proper account of spatial interaction effects.

The past couple of decades have seen what amounts to an intellectual revolution in urban and regional economic research concerning the role of skills in economic growth. From industrial location theory and Alfred Marshall’s concern for agglomeration to more recent research on high-tech districts and industrial clusters, firms and industries have been traditionally the dominant unit of analysis. But since the 1990s, there has been a growing focus on skills. This broad research thrust includes studies of human capital; the creative class and occupational class more broadly; and physical, cognitive, and social skills, among others. This research highlights the growing geographic divergence of skills across cities and metros and their effects on regional innovation, wages, incomes, and development broadly. An expanding literature notes the growing importance of place in organizing and mobilizing these skills. Studies have focused on the role of amenities, universities, diversity, and other place-related factors in accounting for the growing divergence of skills across locations. This chapter summarizes the key lines of research that constitute the skills revolution in urban and regional research.

This chapter outlines two models for analyzing the relationship between infrastructure and regional growth and discusses relevant empirical examples. The first model adopts a standard spatial equilibrium approach and shows that the effect of new infrastructure on regional activity depends on its direct impacts on local productivity, local amenities, and the price of non-traded goods, especially housing. These impacts are determined, in part, by how existing characteristics of the region complement the specific investment. If infrastructure contributes positively to real amenity-adjusted net wages, the local region increases its attractiveness, and the result is an influx of firms and individuals to the region. In turn, this has dynamic effects that may amplify or attenuate the initial growth impetus. It is also possible that an infrastructure project contributes negatively to real amenity-adjusted net wages, imparting a negative influence on equilibrium regional activity. The second model treats a major infrastructure investment as a real option that gives private sector developers the option, but not the obligation, for further development. The value of this option must be included by authorities when assessing benefits of a new infrastructure project. They need to judge the direct private sector responses to an investment plus the indirect equilibrium responses under alternative states of nature. The model shows that for a major infrastructure project, as in the case of other real options, a certainty equivalent approach is generally inadequate for investment analysis since that approach may underestimate the benefit of a new project when future states are uncertain, learning occurs, and decision-making is sequential.

In spite of the ongoing efforts that several countries made into promoting a more balanced economic development within their territory, economic growth theory and even empirical evidence do not come to a unanimous conclusion on the efficiency of public intervention. As such, this chapter reviews the various strands of the theoretical literature, analyzes the results of empirical estimations in Europe and in the USA where regional development policies are already well established, and provides recommendations for future research in this field.

This chapter surveys the topic of the geography of innovation – not the economics of innovation – and asks several questions: What is innovation? Who innovates? Where do they learn to innovate? The research focus has shifted from innovation and technology to the broader issues of knowledge and innovative capability. The empirical literature has been much narrower in scope, previously focusing on research and development (R&D) and now rarely looking beyond patents. The chapter surveys a broader set of innovation indicators – inputs, outputs, and hidden innovation, much of which is uncovered in large-scale surveys. Empirically, there is a global shift in innovative capability toward Asia, primarily in R&D (but less so in basic research) and in process innovation related to manufacturing. The overall pattern is one of persistent spatial concentration. As a result, a thriving business has emerged to craft policies to enhance innovation and to “construct advantage” in an uncertain competitive landscape. Finally, the actors in innovation include not only individual scientists and inventors but also the organizations that employ them, such as universities and firms. It is entrepreneurs who largely determine how innovation is exploited. The fruitful concept of the knowledge filter and the role of entrepreneurship and the geography of entrepreneurship provide clues to the patterns seen.

Generation and diffusion of innovation are two distinct processes that are interlinked in several ways. First, innovation efforts of firms are stimulated by the diffusion of innovation ideas. Second, the market penetration of successful product innovations diffuse to user firms and consumers, providing users opportunities to adopt novel routines and to imitate new designs. Third, creative destruction develops when a novel product finds its way to customers and replaces earlier product vintages, and this phenomenon has the nature of a substitution process. All these processes are supported by knowledge flows which vary in intensity and diversity across the innovation milieu of functional regions. It is concluded that the milieu characteristics which stimulate innovation also stimulate adoption of novelties.

New knowledge generated by an economic agent in a region will tend over time to flow to other economic agents in the same region but also to economic agents in other regions. It is quite common in the literature to use the concept of knowledge spillovers for such knowledge flows, irrespective of whether they are intended or non-intended. The potential for intra-regional knowledge spillover effects depends on the volume and character of the generation on new knowledge in each region as well as of the general characteristics of the individual regional economic milieu, that is, those location attributes, which are regionally trapped and which include how well integrated it is with other regions. The larger this potential, the higher the probability that firms dependent upon knowledge spillovers will locate there and the higher the probability that entrepreneurs will take advantage of this potential to launch innovations and to create new knowledge-based firms. To the extent that firms and entrepreneurs can enjoy these knowledge spillovers, they represent an externality or more specifically a knowledge externality in the regional economy.Great importance is in the literature attributed to knowledge spillovers and knowledge externalities as drivers of regional economic development. Some authors, for example, claim that regional variations in localized knowledge spillovers are one of the main reasons behind regional variations in innovation performance. Against this background, the purpose of this chapter is, based upon a general characterization of knowledge flows, to analyze the character of knowledge externalities and, in particular, their sources, their economic nature, their recipients, their mechanisms and channels, their geographic reach, and their economic consequences generally and for regional economic development in particular.

Over the last three decades, literature on industrial districts, innovative milieus, and industrial clusters has enriched our knowledge about endogenous factors and processes driving regional development and the role of the region as an important level of economic coordination. This class of stylized development concepts has emerged since the 1970s and attempts to account for successful regional adaptations to changes in the global economic environment. Each of these concepts grew out of specific inquiries into the causes of economic success to be found in the midst of general decline by building upon the early ideas of Alfred Marshall in several ways. Neo-Marshallian districts found in Italy highlight the importance of small firms supported by strong family and local ties, while the innovative milieu concept places great emphasis on the network structure of institutions to diffuse externally sourced innovations to the local economy. Clusters have become far more general in scope, fruitful in theoretical insights, and robust in application, informing the work of both academics and policy-makers around the world.

In this chapter, an overview is presented of the three-phase evolution thus far of the regional systems of innovation perspective. The connected notion of the “learning region” is situated and subsequently re-situated in this account. The chapter begins by establishing the debate in the regional governance, learning, and policy contexts, especially with reference to the concept of “experimental regionalism.” Early reflections upon various critical responses to the 20-year literature on regional innovation represent the first main phase change, indicating the relative conceptual and empirical flexibility of the approach. Innovation in thinking about entrepreneurship is shown to have been at the heart of this first evolving perspective on regional dynamics. The most recent phase change represents the engagement of regional innovation systems, as a core subfield of evolutionary economic geography, with key concepts in the complexity sciences. These are coevolution, complexity, and emergence. Each is shown to denote important new ways of thinking about regional innovation and evolution. The continuing relevance of the perspective to regional theoretical and policy application is underscored.

This chapter provides an overview of current theories and empirical research on cities and the knowledge economy. Two recent and interrelated streams of literature are discussed: the first focusing on agglomeration economies related to increasing returns and knowledge spillovers of firms in cities and the second highlighting the role of knowledge workers and creativity in identifying new and innovative growth opportunities in cities. We argue that analyses using knowledge production functions to capture knowledge flows in cities do not, as of yet, provide true insight into the generation and transfer of different kinds of knowledge. Only recently are various conceptualizations of distance and knowledge transmission channels able to address the heterogeneity of the actors and processes involved in capturing the respective role of cities in knowledge creation. We conclude that the mechanisms that create and diffuse knowledge in cities should be better embedded into both streams of literature. The current discourse on agglomeration externalities obviously needs such conceptual and methodological views to address current impasses. In particular, evolutionary economic geographical concepts are promising in explaining the innovative behavior of growing firms and organizations in cities, carefully addressing the heterogeneity of the actors involved, spatial scale, selection and survival, as well as time and path dependency.

This chapter reviews the importance of networks in the innovation process from a spatial perspective. Such networks are part of different scale systems of innovation and are essential to the creation of knowledge externalities. It is well established in the extant literature that innovation does not occur in isolation, and furthermore, interorganizational networks facilitate innovation creation. Social networks, trust, and local embeddedness play key roles in the formation of such networks. In addition, relational perspectives, such as non-geographical proximities, are also vital factors for the creation of innovation networks, the main objective of which is knowledge creation. Important enough, the latter can be approached as crucial production factor in the frame of the knowledge economy. Moreover, scale is an important attribute of such networks, as both local and global links are important in the innovation process.

Within location theory, classical models are typical abstract and formalized models, in which the main reasoning behind location choice of firms is driven by the minimization of transportation costs to achieve natural and intermediate production resources, and markets for final goods that are territorially dispersed. Classical models are similar in the question they want to reply to: what economic logic explains the location choices of firms in space? This topic is an important one. Although in terms of time and financial resources, the performance of transport and communication has improved enormously, many economic activities have not become footloose to the extent expressed by the “death of distance.” Their location choice still remains anchored to a balance between a physical location generating economic advantages – in the form of agglomeration economies – and transport costs to intermediate or final markets, as explained by these models.

This chapter reviews some of major thematic approaches which have characterized urban and regional research over recent decades. Three broad schools of research are discussed, namely, the new economic geography, the new urban agenda, and the evolutionary and institutional school. The major assumptions underlying each of the schools of thought are outlined, and the broad areas of agreement and disagreement between the three schools of thought are highlighted. The changing economic realities on the ground in many regions, whereby the previously dominant large cities are no longer the key drivers of economic growth, pose major conceptual, analytical, and empirical challenges to all three of these schools of thought, schools which had emerged precisely during the period when major cities were reemerging as the drivers of growth.

This chapter does not aim to survey what has been accomplished in new economic geography (NEG) since the publication of Paul Krugman’s seminal paper. Rather, we provide an overview of recent developments in the NEG literature that build on the idea that the difference in the economic performance of regions is explained by the behavior and interactions between households and firms located within them. This means that we consider NEG models which take into account land markets, thereby the internal structure and industrial mix of urban agglomerations.

In this chapter we first briefly discuss how the new economic geography literature (NEG) follows from and builds on international trade theory. We then turn to the main empirical implications of NEG. We highlight that the main problem with empirical applications of NEG is that a single test of the implications of the model combined is illusive because of the structure of the model. As a result the main consequences of the model are usually tested separately. And some of the implications of the model are also consistent with other models. We stress, therefore, that despite a real surge in empirical NEG inspired research, the empirical evidence is still rather sketchy and also that so far NEG-based policy advice is still mostly qualitative.

In the past decade, economic geography has encountered increasing interest and debates about evolutionary and relational thinking in regional development. Rather than comparing the two approaches, this chapter investigates how they can complement one another and be applied to specific research fields in economic geography. A comparison would be difficult because the approaches address different levels of the research process and are in a relatively early stage of their development. To demonstrate the potential of combining the two approaches, this chapter aims to conceptualize cluster dynamics in an integrated relational-evolutionary perspective. In recent years, research on clusters has experienced a paradigmatic shift from understanding their network structure to analyzing dynamic changes. Within this context, inspired by relational and evolutionary thinking, a comprehensive tripolar analytical framework of cluster evolution is developed that combines the three concepts of context, network, and action, allowing each to evolve in interaction with the others. Through this, the chapter argues that, rather than viewing relational and evolutionary accounts as competitive approaches to economic geography, they can, in an integrated form, become fundamental guides to economic geography research.

The concept of path dependence has rapidly assumed the status of a “fundamental principle” in the new paradigm of evolutionary economic geography that has emerged over the past few years. This chapter reviews the interpretation and use of this concept within this new field. The dominant interpretation has been that of “lock-in,” by self-reinforcing mechanisms, of particular (equilibrium) patterns of industrial location and regional specialization. This model is somewhat restrictive, however, and does not capture the full repertoire of ongoing path-dependent evolutionary trajectories that can be observed in the economic landscape. To respond to this limitation, the chapter suggests a “developmental–evolutionary” model of path dependence that includes “lock-in” as a special case, but which is also more general in its application and relevance.

This chapter discusses the literature on agglomeration economies from the perspective of jobs and job dynamics. It provides a partial review of the empirical evidence on agglomeration externalities; the functionality of cities; the dynamic relationship between cities, jobs, and firms; and the linkages between cities. We provide the following conclusions. First, agglomeration effects are quantitatively important and pervasive. Second, the productive advantage of large cities is constantly eroded and needs to be sustained by new job creations and innovations. Third, this process of creative destruction in cities, which is fundamental for aggregate growth, is determined by the characteristics of urban systems and broader institutional features. We highlight important differences between developing countries and more advanced economies. A major challenge for developing countries is the transformation of their urban systems into drivers of economic growth.

This chapter looks at the recent developments in economic geography theory and sets out to shed light on its contribution to the understanding of the dynamics of technological change. The replacement of the linear model with more sophisticated conceptualizations of the process of innovation has made it possible to account for persistent disparities in innovative performance across space and has motivated researchers to incorporate the role of space and places in the analysis of innovation processes. From the physical-metrical approach of geography as distance to the emphasis on specialization and diversification patterns (geography as economic place), institutional-relational factors, nonspatial proximities, and “integrated” frameworks, economic geography theory has substantially evolved in terms of its contribution to the understanding of technological dynamics with significant implications for the rationale, design, and implementation of innovation policies.

This chapter is concerned with the process by which geographical economics influences policy. It considers a number of barriers that limit this influence focusing specifically on the availability of data, the limitations of spatial analysis, and the role of the evaluation of government policy. It considers why these problems present such significant barriers and proposes some solutions. In terms of the availability of data, the chapter explains why problems concerning the correct unit of analysis and measurement error may be particularly acute for spatial data (especially at smaller spatial scales). Resulting concerns about the representativeness of data and the mismatch between functional and administrative units may further hamper interaction with policy makers. For spatial analysis, the major problem concerns the extent to which empirical work identifies the causal factors driving spatial economic phenomena. It is suggested that greater focus on evaluating the impact of policies may provide one solution to this general identification problem.

This chapter focuses on the ways in which travel behavior and demand are analyzed within the framework of regional science. Unlike numerous recent surveys that cover the more technical and abstract aspects of mathematically modeling travel behavior and demand, the attention here is more on the practical aspect of applying travel behavior and demand analysis to subjects such as regional development, infrastructure investment, and congestion analysis. Thus, while the main methods of modeling travel behavior and demand are outlined and critiqued, there is also considerable references to such things as demand elasticities and their estimation that are at the core of applied regional analysis. These types of parameter provide a direct link between a soft policy shift or a harder infrastructure investment, travel behavior, and ultimately the implications of this for regions. There is also discussion of the uses made of the forecasts that are the de facto rationale for studying travel behavior and travel demand, and the ways that neutral forecasting can be manipulated in decision-making.

Activity-based analysis (ABA) is an approach to understanding transportation, communication, urban, and related social and physical systems using individual actions in space and time as the basis. Although the conceptual foundations, theory, and methodology have a long tradition, until recently an aggregate trip-based approach dominated transportation science and planning. Changes in the business and policy environment for transportation and the increasingly availability of disaggregate mobility data have led to ABA emerging as the dominant approach. This chapter reviews the ABA conceptual foundations and methodologies. ABA techniques include data-driven methods that analyze mobility data directly as well as develop inputs for ABA modeling. ABA models include econometric models, rule-based models and microsimulation/agent-based models. This chapter concludes by identifying major research frontiers in ABA.

This chapter begins with a discussion of how communications technologies have reduced the influence of distance on the location of economic activity. The origins of network analysis in regional science are described. The importance of social networks and social network science in sociology and related disciplines during the 1970s, 1980s, and 1990s is explained. This is followed by a discussion of new discoveries concerning the structure of the Internet that took place in the late 1990s. The rise of social media, the continued development of social network science, and the popularity of social network sites such as Facebook, Twitter, and LinkedIn in the new millennium are then depicted along with the most recent research findings that derive from connectivity and contagion processes within social networks. The chapter concludes with an account of methods for determining the importance of distance in influencing social and economic activity in the new world of social networks.

The relationship between urban development and transport is not simple and one way but complex and two way and is closely linked to other urban processes, such as macroeconomic development, interregional migration, demography, household formation, and technological innovation. In this chapter, one segment of this complex relationship is discussed: the two-way interaction between urban land use and transport within urban regions. The chapter looks at integrated models of urban land use and transport, i.e., models that explicitly model the two-way interaction between land use and transport to forecast the likely impacts of land use and transport policies for decision support in urban planning. The discussion starts with a review of the main theories of land-use transport interaction from transport planning, urban economics, and social geography. It then gives a brief overview of selected current operational urban models, thereby distinguishing between spatial-interaction location models and accessibility-based location models, and discusses their advantages and problems. Next, it reports on two important current debates about model design: are equilibrium models or dynamic models preferable, and what is the most appropriate level of spatial resolution and substantive disaggregation? This chapter closes with a reflection of new challenges for integrated urban models likely to come up in the future.

Methods for the analysis and prediction of travel conforming to macroscopic assumptions about choices of the urban population cut a broad swath through the field of regional science: economic behavior, spatial analysis, optimization methods, parameter estimation techniques, computational algorithms, network equilibria, and plan evaluation and analysis. This chapter seeks to expose one approach to the construction of models of urban travel choices and implicitly location choices. Beginning with the simple route choice problem faced by vehicle operators in a congested urban road network, exogenous constants are relaxed and replaced with additional assumptions and fewer constants, leading toward a more general forecasting method. The approach, and examples based upon it, reflects the author’s research experience of 40 years with the formulation, implementation, and solution of such models.

We overview some of the major advances in supply chains and transportation networks, with a focus on their common theoretical frameworks and underlying behavioral principles. We emphasize that the foundations of supply chains as network systems can be found in the regional science and spatial economics literature. In addition, transportation network concepts, models, and accompanying methodologies have enabled the advancement of supply chain network models from a system-wide and holistic perspective.We discuss how the concepts of system optimization and user optimization have underpinned transportation network models and how they have evolved to enable the formulation of supply chain network problems operating (and managed) under centralized or decentralized, that is, competitive, decision-making behavior.We highlighted some of the principal methodologies, including variational inequality theory, that have enabled the development of advanced transportation network equilibrium models as well as supply chain network equilibrium models.

The modern spatial economy has a global “networked” character that is generating important socioeconomic and political changes. In this respect, new forms of connectivity play a significant role through their dynamic and complex interplay with the economic and political driving forces behind globalization. In analyzing such impacts, it is useful to consider the tools and models that have been adopted in regional economics as well as in other disciplines. In this context, it is also necessary to reflect on complexity theory and on the models able to map out the complex interconnected spatial networks.This chapter begins with a concise review of the most important definitions of complexity, in the light of their relations with spatial networks. There follows an exploration of the main findings from two “close” disciplines, that is, spatial economics and network science, with reference to their associated approaches and modeling tools which are able to grasp complexity from, respectively, the behavioral and the network structure viewpoint. The emerging discussion – with reference to both static and dynamic frameworks through the lens of complexity issues – indicates that (i) a formal correspondence between the fundamental spatial economic models and network models exists and (ii) this correspondence highlights the “simplicity” of the laws underlying complex spatial networks.

Location theory has traditionally been based on equilibrium concepts. Dynamics have been introduced mainly to ascertain whether there are paths leading to the equilibrium states. The modeling of dynamics has been very simple yet involving both locational changes and price changes. Notions of market areas and competition between firms have been at the core of location analysis. Although the classical location theory was developed in a regional context, the models have found a number of recent applications in urban analysis where interdependencies and dynamics are central elements. The theoretical contributions of Hotelling, Hoover, and Palander form cornerstones for the discussion in the current chapter. In this chapter, we will mainly dwell in the Hotelling tradition and use the theories of Hoover and Palander as introductory and complementary inputs. The chapter presents a series of behavioral models in the spirit of the classical Hotelling location game involving the spatial location of suppliers (sellers) and consumers (customers) in an urban context. The models have been established within a cellular automata framework. The location models studied assume fixed prices. The location of sellers is determined by the relative accessibility to customers and the competition between sellers for customers. Using the techniques of cellular automata, a set of simulations will be performed to discuss equilibrium states of customer-seller systems. The discussion will serve to illustrate some elements of location theory under different levels of complexity.

This chapter introduces into the theory of labor and capital movements between regions or countries. Movements of other mobile factors, in particular knowledge, are not dealt. After an introduction defining terms, it explains the basic factor mobility model assuming perfect competition and full factor price flexibility. Particular emphasis is given to the welfare results: Who are the winners and losers if factors are allowed to move and under what conditions does free mobility increase overall efficiency? We show how factor allocations deviate from an efficient outcome if the markets do not work perfectly. After studying factor mobility in a static framework, we extend the analysis to a dynamic framework. It is needed because investment decisions are forward looking. Investors compare present expenditures with present values of future returns. The same holds true for migration because migrants invest into human capital when they expend migration cost today in order to earn a higher income in the future. In a final section, we study the role of factor mobility in New Economic Geography. A concluding section points to further topics not dealt with in this chapter.

This chapter presents and critically evaluates the economic assumptions and applicability of a series of regional and interregional interindustry models. It begins with the demand-driven, single-region Leontief quantity model and its cost-push price dual. Then Section 45.4 discusses the ideal, full information, interregional input–output model with interregional spillover and feedback effects at length, and compares it with the requirements and assumptions of more limited information, multiregional input–output models. Section 45.5 discusses how to construct and add an interregional consumption function to obtain the type II interregional interindustry model. Section 45.6 outlines further extensions, all through to the most complex price-quantity interacting interregional demo-economic model LINE. Finally, an Appendix presents the microeconomic foundation for the Leontief model and compares it with the alternative supply-driven quantity model and its demand-pull price dual.

Interregional trade has been relatively neglected by most trade analysts. A dearth of data has limited formal explorations of interregional trade but the magnitudes of the volumes revealed suggest that greater attention should be directed to this form of connectivity between economies. This chapter begins with a review of the theory and practice of international trade theory and its link to some of the ideas that form the basis of the New Economic Geography. Some alternative approaches to the measurement of trade are examined, especially the role of intra-industry as opposed to interindustry trade, vertical specialization, trade overlap, and spatial production cycles. Thereafter, attention is addressed to the interregional impacts of international trade.

Uncertainty affects the dynamic trade-offs of environmental and natural resource management in a variety of ways and forms. The uncertain responses to anthropogenic activities may be due to genuine stochastic processes that drive the evolution of the underlying natural systems or simply due to our poor understanding of these complex systems and their interactions with the exploitation policies. These interactions are of particular importance when the ecosystem response might involve irreversibility, so that unexpected undesirable outcomes cannot be undone after they are realized. In this chapter, we review the various sources of uncertainty, the methodologies developed to account for them, and the implications regarding the management of environmental and natural resources.

We cover applications of game theory in environmental and resource economics with a particular emphasis on noncooperative transboundary pollution and resource games. Both flow and stock pollutants are considered. Equilibrium concepts in static and dynamic games are reviewed. We present an application of game theoretical tools related to the formation and sustainability of cooperation in transboundary pollution games. We discuss the analytical tools relevant for the case of a stock pollutant and offer an application related to the optimal institutional arrangement to regulate a pollutant when several jurisdictions are involved.

Commensurate valuation of market and nonmarket public goods allows for a more valid benefit-cost analysis. Economic methods for valuing nonmarket public goods include actual behavior-based revealed preference methods such as the hedonic property method for urban-suburban public goods and travel cost-based models for outdoor recreation. For valuing proposed public goods for which there is no current behavior or valuing the existence or passive use values of public goods, economists can rely upon stated preference methods. While there is skepticism among some economists for relying upon what people say they will pay rather than what their actual behavior suggests they will pay, there is general acceptance of stated preference methods. These stated preference methods include the well-known contingent valuation method and choice experiments (sometimes called conjoint analysis). Lastly, in situations where there is neither time nor money to conduct an original revealed or stated preference study, economists typically rely upon benefit transfers from existing revealed preference and stated preference studies to provide rough estimates of the values of public goods such as water quality, air quality, wetlands, recreation, and endangered species.

Benefit-cost analysts attempt to compare two states of the world, the status quo and a state in which a policy having benefits and costs is being contemplated. For environmental policies, this comparison is greatly complicated by the difficulty in inferring the values that individuals place on an increment to environmental quality. Unlike ordinary private goods, environmental goods are not directly exchanged in markets with observable prices. In this chapter, the hedonic approach to inferring the benefits of an environmental policy is examined.

This chapter presents an overview of the mass balance principle and its applications. It is an important tool for quantifying wastes which are produced by economic processes. These wastes are equal in mass to the difference between total raw material inputs to the process and useful material outputs. Products are becoming more complex which results in an increase of input mass and wastes. It is safe to say that nowadays process wastes far exceed the mass of materials that are finally embodied in useful products.The application of the mass balance principle can take many shapes and forms, and this chapter illustrates a few. Using mass balance and chemical engineering knowledge of processes, we found that on a yearly basis, the inorganic chemical industry has a yield of 91 % (9 % of the inputs end up in waste), and the organic chemical industry has a yield of 40 %. A second example is the rare earth metal industry, where potential recovery of these scarce metals is quantified to motivate reuse and recycling. Presently less than 1 % of rare earth metals are recovered from end-of-life products, but as the demand for these resources increases in the near future for products such as electric motors and wind power turbines, recovery will become necessary.An introduction to thermodynamics and exergy is included, since all wastes are thermodynamically degraded as compared to raw materials. The exergy of the inputs, products, and wastes is an important factor to consider for process efficiency and environmental evaluation.

Environmental and natural resource economics has long wrestled with spatial elements of human behavior, biophysical systems, and policy design. The treatment of space by academic environmental economists has evolved in important ways over time, moving from simple distance measures to more complex models of spatial processes. This chapter presents knowledge developed in several areas of research in spatial environmental and natural resource economics. First, it discusses the role played by spatial heterogeneity in designing optimal land conservation policies and efficient incentive policies to control pollution. Second, it describes the role space plays in nonmarket valuation techniques, especially the hedonic and travel cost approaches which inherently use space as a means to identify values of nonmarket goods. Third, it explains a set of quasi- or natural-experimental empirical methods which use spatial shocks to estimate the effects of pollution or environmental policy on a wide range of outcomes such as human health, employment, firm location decisions, and deforestation. Finally, it describes spatial models of human behavior including locational sorting and the interaction of multiple agents in a land use/conservation setting. The chapter ends with a discussion of some promising future areas for further evolution of the modeling of space in environmental economics.

The expected global impacts of climate change can be attributed to a set of common stressors. The magnitude of specific impacts, however, depends on the extent to which regional resources – from ecosystems to human-made infrastructures – are at risk and the abilities of regions to mitigate that risk. This chapter begins with an overview of some of the impacts expected from climate change, stratified by the density of populations and economic activities. Then we review differences in risk and mitigation capacities across major regions. The inherent interconnection of environmental, economic, and social dimensions of climate impacts underscores the need to assess climate change impacts in ways that address these dimensions.

Sustainability has become a key concept in the quest to define a normative framework for urban and regional development. This chapter presents an overview of what is meant by sustainability first from the regional and then from the city level. Both scales have a long history in the planning domain, but the notion of a sustainable city is key to both realms and is the main focus of this chapter. While there is widespread agreement on broad parameters and principles about urban and regional sustainability, there are entrenched debates over implementation. On one level, there are debates over implementation methods, especially the degree to which partial success in implementation is better or worse than doing nothing. More fundamental debates about sustainability involve the distinction between process vs. form and the integration of city versus nature.

The impact of human population growth on the environment represents the major challenge of our time. This chapter reviews demographic change over the last century, set in historical context, and different perspectives on population-environment interactions. Differences in population growth rates and demographic change across space are explored, followed by perspectives on the population-environment nexus at multiple scales with a particular focus on those contexts where impacts are likely to be the very greatest on humankind. The alignment of individual and higher-level actions resulting in environmental impacts and the negative force of the impacts of actions are important, to signal the need to change behaviors. Interrelationships are shown to be highly complex. It is argued that multidisciplinary efforts to tackle complexitycomplexity and to focus on resilienceresilience at multiple scales are critically needed, with the importance of multidisciplinary regional science thought being underscored. The question is raised, however, whether these efforts will be coordinated well enough across multiple scales and with multiple disciplines and publics to avoid what could be catastrophic impacts. These are most likely to occur at local and regional scales where population growth rates are high, natural environments already vulnerable, and resilience limited.

This chapter begins with definitions of geographic information science (GIScience), of geocomputation, and of spatial analysis. We then discuss how these research areas have been influenced by recent developments in computing and data-intensive analysis, before setting out their core organizing principles from a practical perspective. The following section reflects on the key characteristics of geographic information, the problems posed by large data volumes, the relevance of geographic scale, the remit of geographic simulation, and the key achievements of GIScience and geocomputation to date. Our subsequent review of changing scientific practices and the changing problems facing scientists addresses developments in high-performance computing, heightened awareness of the social context of GIS, and the importance of neogeography in providing new data sources and in driving the need for new techniques.

This chapter provides an introduction to geocomputation and geocomputational methods. As such it considers the scope of the term geocomputation, the principal techniques that are applied, and some of the key underlying principles and issues. Chapters elsewhere in this major reference work examine many of these ideas and methods in greater detail. In this connection it is reasonable to ask whether all of modern spatial analysis is inherently geocomputational; the answer is without doubt “no,” but its growing importance in the development of new forms of spatial analysis, in exploration of the behavior and dynamics of complex systems, in the analysis of large datasets, in optimization problems, and in model validation remains indisputable.

The current ubiquity of data collection is providing unprecedented opportunities for knowledge discovery and extraction. Data sources can be large, complex, heterogeneous, structured, and unstructured. In order to explore such data and exploit opportunities within the data deluge, tools and techniques are being developed to help data users generate hypotheses, explore data trends and ultimately develop insights and formulate narratives with their data. These tools often rely on visual representations of the data coupled with interactive computer interfaces to aid the exploration and analysis process. Such representations fall under the purview of visualization, in which scientists have worked on systematically exploiting the human visual system as a key part of data analysis. Research in this area has been inspired by a number of historical sources, examples include physicist James Maxwell’s sculpture of a thermodynamic surface in 1874, Leonardo da Vinci’s hand-drawn illustration of water from his studies to determine the processes underlying water flow, or the flow map of Napoleon’s March on Moscow produced by Charles Minard in 1869. Each of these examples attempts to explain data in a visual manner, and, as visualization has progressed, principles and practices have been adopted to standardize representations, and, more importantly, better exploit properties of the human visual system.

The modifiable areal unit problem (MAUP) is a serious analytical issue for analysts using spatial data. The MAUP manifests itself through the instability of a wide range of statistical results derived from analysis on spatially organized data. When spatial data are aggregated, the results are conditional on the spatial scale at which they are conducted, and the configuration of the areal units that are employed to represent the data. Such uncertainty means that the results of spatial data where the MAUP has not been considered explicitly should be treated with caution. Although solutions have been proposed, none have been applicable in more than a couple of specific cases. As such, it is likely that the MAUP will never be truly solved. This chapter charts the two related aspects of the MAUP, the scale and zonation effects, and details the role of spatial autocorrelation in understanding the processes in the data that lead to the statistical nonstationarity. The role of zone design as a tool to enhance analysis is explored and reference made to analyses that have adopted explicit spatial frameworks.

As the number, volume and resolution of spatio-temporal datasets increases, traditional statistical methods for dealing with such data are becoming overwhelmed. Nevertheless, the spatio-temporal data are rich sources of information and knowledge, waiting to be discovered. The field of spatio-temporal data mining (STDM) emerged out of a need to create effective and efficient techniques in order to turn the massive data into meaningful information and knowledge. This chapter reviews the state of the art in STDM research and applications, with emphasis placed on three key areas, including spatio-temporal prediction and forecasting, spatio-temporal clustering and spatio-temporal visualization. The future direction and research challenges of STDM are discussed at the end of this chapter.

This chapter outlines the key ideas of Bayesian spatial data analysis, together with some practical examples. An introduction to the general ideas of Bayesian inference is given, and in particular the key rôle of MCMC approaches is emphasized. Following this, techniques are discussed for three key types of spatial data: point data, point-based measurement data, and area data. For each of these, examples of appropriate kinds of spatial data are considered and examples of their use are also provided. The chapter concludes with a discussion of the advantages that Bayesian spatial analysis has to offer as well as considering some of the challenges that this relatively new approach is faced with.

Two classes of models that have made major breakthroughs in regional science in the last two decades are cellular automata (CA) and agent-based models (ABM). These are both complex systems approaches and are built on creating microscale elemental agents and actions that, when permuted over time and in space, result in forms of aggregate behavior that are not achievable by other forms of modeling. For each type of model, the origins are explored, as are the key contributions and applications of the models and the software used. While CA and ABM share a heritage in complexity science and many properties, nevertheless each has its own most suitable application domains. Some practical examples of each model type are listed and key further information sources referenced. In spite of issues of data input, calibration, and validation, both modeling methods have significantly advanced the role of modeling and simulation in geography and regional science and gone a long way toward making models more accountable and more meaningful at the base level.

Spatial microsimulation is an excellent option to create estimated populations at a range of spatial scales where data may be otherwise unavailable. In this chapter, we outline three common methods of spatial microsimulation, identifying the relative strengths and weaknesses of each approach. We conclude with a worked example using deterministic reweighting to estimate tobacco smoking prevalence by neighborhood in London, UK. This illustrates how spatial microsimulation may be used to estimate not only populations but also behaviors and how this information may then be used to predict the outcomes of policy change at the local level.

Spatial networks organize and structure human social, economic, and cultural systems. The analysis of network structure depends on the development of measures and models of networks, which in turn rely on mathematical graph theory. Key concepts and definitions from graph theory are reviewed and used to develop a variety of graph structural measures, which can be used to investigate local and global network structure. Particular emphasis is placed on high-level network structural features of centrality, cohesive subgraphs, and structural equivalence. Widely used models for spatial networks are introduced and discussed. Pointers to empirical research on real-world spatial networks are provided.

We review some of the special properties of spatial data and the ways in which these have influenced developments in spatial data analysis. We adopt a historical perspective beginning in the early twentieth century before moving to the development of spatial autocorrelation statistics in geography’s Quantitative Revolution. Phases of development after the Quantitative Revolution are divided into emergence of spatial econometrics, the development of exploratory methods for spatial data analysis, and local statistics for handling heterogeneity. We then consider more recent advances in the areas of spatial data mining, the “new” geostatistics, and Bayesian hierarchical statistical modeling of spatial data.

In this chapter, we discuss key concepts for exploratory spatial data analysis (ESDA). We start with its close relationship to exploratory data analysis (EDA) and introduce different types of spatial data. Then, we discuss how to explore spatial data via different types of maps and via linking and brushing. A key technique for ESDA is local indicators of spatial association (LISA). ESDA needs to be supported by software. We discuss two main lines of software developments: GIS-based solutions and stand-alone solutions.

Spatial autocorrelation in health events may be the signature of underlying causal factors of direct scientific and practical interest but may also be due to pedestrian or nuisance factors that obscure meaningful spatial patterns. The problem is to discern spatial patterns that inform our understanding of the health events themselves from those that are of little interest. This chapter provides a framework for advancing knowledge when the causes of observed health event clusters are unknown.

Describing area-based differences in health outcomes has a long history (Kawachi and Berkman Neighborhoods and health. Oxford University Press, Oxford, 2003), and evidence of ecologic variations in health sparks interest from multiple perspectives. In particular, researchers investigate these ecologic variations for surveillance and monitoring of health disparities (Krieger et al. Am J Public Health 95:312–323, 2005) and to understand the impacts contexts have on individuals. In the latter category, causal questions motivated by evidence of area-based differences in health include the following: How much do contexts, such as neighborhoods, impact health? What is the impact of a specific contextual exposure on health? How do contexts mediate the effects of individual-level health risk factors? Ecologic factors may have tremendous importance for population health (Kawachi and Berkman Neighborhoods and health. Oxford University Press, Oxford, 2003), underscoring the value of recognizing opportunities and methodological challenges for causal inference when ecological variations in health are present. We address these issues as follows: we begin by identifying what constitutes a multilevel data analysis and present a discussion on how a range of data structures that are observed in the real world, or due to sampling design, can be accommodated within a multilevel framework. We discuss the types of research questions that typically motivate multilevel analyses and contrast the application of multilevel methods against other approaches for answering such questions with an emphasis on causal inference. After laying down the substantive motivation to utilize multilevel methods, key statistical models are specified with a description of the properties of each model. We close by presenting extensions to the basic multilevel model that allow us to incorporate realistic complexity in our analyses.

This chapter provides an overview of spatial dynamics in the field of regional science. After defining the context of spatial dynamics and the alternative conceptualizations of space and time, the chapter surveys the various areas of substantive interest where spatial dynamics come to the fore. A second focus is on the methodological and technical issues surrounding the methods of space-time data analysis. Here the emphasis is on exploratory methods for space-time data focusing on the evolution of spatial patterns as well as the identification of temporal dynamics that cluster in space.

Spatial sampling is the process of collecting observations in a two-dimensional framework. Careful attention is paid to (1) the quantity of the samples, dictated by the budget at hand, and (2) the location of the samples. A sampling scheme is generally designed to maximize the probability of capturing the spatial variation of the variable under study. Once initial samples have been collected and its variation documented, additional measurements can be taken at other locations. This approach is known as second-phase sampling, and various optimization criteria have recently been proposed to determine the optimal location of these new observations. In this chapter, we review fundamentals of spatial sampling and second-phase designs. Their characteristics and merits under different situations are discussed, while a numerical example illustrates a modeling strategy to use covariate information in guiding the location of new samples. The chapter ends with a discussion on heuristic methods to accelerate the search procedure.

Bayesian inference has been at the center of the development of spatial statistics in recent years. In particular, Bayesian hierarchical models including several fixed and random effects have become very popular in many different fields. Given that inference on these models is seldom available in closed form, model fitting is usually based on simulation methods such as Markov chain Monte Carlo.However, these methods are often very computationally expensive and a number of approximations have been developed. The integrated nested Laplace approximation (INLA) provides a general approach to computing the posterior marginals of the parameters in the model. INLA focuses on latent Gaussian models, but this is a class of methods wide enough to tackle a large number of problems in spatial statistics.In this chapter, we describe the main advantages of the integrated nested Laplace approximation. Applications to many different problems in spatial statistics will be discussed as well.

Spatial statistics has in the last decade or two emerged as a major sub-specialism within statistics. Applications areas are diverse, and there is cross-fertilization with methodologies in other disciplines (econometrics, epidemiology, geography, geology, climatology, ecology, etc). This chapter reviews three major settings and techniques that have attracted attention from statisticians: spatial econometrics and simultaneous autoregressive models, spatial epidemiology and conditional autoregressive models, and geostatistical methods for point pattern data. The review is oriented to Bayesian inferences for such models, including discussion of choice of prior densities, questions of identification, outcomes of interest, and methods of estimation (using Markov chain Monte Carlo).

Geographically weighted regression (GWR) was proposed in the geography literature to allow relationships in a regression model to vary over space. In contrast to traditional linear regression models, which have constant regression coefficients over space, regression coefficients are estimated locally at spatially referenced data points with GWR. The motivation for the introduction of GWR is the idea that a set of constant regression coefficients cannot adequately capture spatially varying relationships between covariates and an outcome variable. GWR is based on the appealing idea from locally weighted regression of estimating local models for curve fitting using subsets of observations centered on a focal point. GWR has been applied widely in diverse fields, such as ecology, forestry, geography, and regional science. At the same time, published work from several researchers has identified methodological issues and concerns with GWR and has questioned the application of the method for inferential analysis. One of the concerns with GWR is with strong correlation in estimated coefficients for multivariate regression terms, which makes interpretation of map patterns for individual terms problematic. The evidence in the literature suggests that GWR is a relatively simple and effective tool for spatial interpolation of an outcome variable and a more problematic tool for inferring spatial processes in regression coefficients. The more complex approach of Bayesian spatially varying coefficient models has been demonstrated to better capture spatial nonstationarity in regression coefficients than GWR and is recommended as an alternative for inferential analysis.

Characterizing the spatial structure of variables in the regional sciences is important for several reasons. Firstly, the spatial structure may itself be of interest. The structure of a population variable tells us something about how the population is configured spatially. For example, is the population clustered by some properties, but not others? Secondly, mapping variables from sparse sample observations or transferring values between areal units requires knowledge of how the property of interest varies spatially. Thirdly, we require knowledge of spatial variation in order to design sampling strategies which make the most of the effort, time, and money expended in sampling. Geostatistics comprises a set of principles and tools which can be applied to characterize or model spatial variation and use that model to optimize the mapping, simulation, and sampling of spatial properties. This chapter provides an introduction to some key ideas in geostatistics, with a particular focus on the kinds of applications which may be of interest for regional scientists.

This chapter provides an introductory discussion of spatial autocorrelation (SA), which refers to correlation existing and observed in geospatial data, and which characterizes data values that are not independent, but rather are tied together in overlapping subsets within a given geographic landscape. This chapter summarizes the various interpretations of SA, one being map pattern. SA can be quantified in a number of different ways, too, one being with the Moran Coefficient. Spatial filtering is a statistical method whose goal is to obtain enhanced and robust results in a spatial data analysis by decomposing a spatial variable into trend, a spatially structured random component (i.e., spatial stochastic signal), and random noise. Its aim is to separate spatially structured random components from both trend and random noise, and, consequently, leads statistical modeling to sounder statistical inference and useful visualization. This separation procedure can involve eigenfunctions of the matrix version of the numerator of the Moran Coefficient. This chapter summarizes the eigenvector spatial filtering (ESF) conceptual material, and presents the computer code for implementing ESF in R, Matlab, MINITAB, FORTRAN, and SAS. Next, it demonstrates that eigenvector spatial filter estimators are unbiased, efficient, and consistent. Finally, it summarizes an ESF empirical example application, and the extension of ESF to spatial interaction modeling.

This chapter provides a selective survey of specification issues in spatial econometrics. We first present the most commonly used spatial specifications in a cross-sectional setting in the form of linear regression models including a spatial lag and/or a spatial error term, heteroscedasticity or parameter instability. Second, we present a set of specification tests that allow checking deviations from a standard, that is, nonspatial, regression model. An important space is devoted to unidirectional, multidirectional, and robust LM tests as they only require the estimation of the model under the null. Because of the complex links between spatial autocorrelation and spatial heterogeneity, we give some attention to the specifications incorporating both aspects and to the associated specification tests.

Past applications of spatial regression models have frequently interpreted the parameter estimates of models that include spatial lags of the dependent variable incorrectly. A discussion of issues surrounding proper interpretation of the estimates from a variety of spatial regression models is undertaken. We rely on scalar summary measures proposed by LeSage and Pace (Introduction to spatial econometrics. Taylor Francis/CRC Press, Boca Raton, 2009) who motivate that these reflect a proper interpretation of the marginal effects for the nonlinear models involving spatial lags of the dependent variable. These nonlinear spatial models are contrasted with linear spatial models, where interpretation is more straightforward. One of the major advantages of spatial regression models is their ability to quantify spatial spillovers. These can be defined as situations where nonzero cross-partial derivatives exist that reflect impacts on outcomes in region i arising from changes in characteristics of region j. Of course, these cross-partial derivatives can be interpreted as impacts of changes in an own region characteristic on other regions or changes in another regions’ characteristic on the own region. The ability to produce empirical estimates along with measures of dispersion that can be used for inference regarding the statistical significance, magnitude, and spatial extent of spillovers provides a major motivation for using spatial regression models.

Maximum likelihood estimation has been the standard method employed for estimating spatial econometric models. This chapter introduces these methods, examines the specific case of a spatial error model, and provides an example based on a large data set. In addition, the chapter sets forth various solutions to the computational difficulties that arise for large data sets.

This chapter provides a survey of the recent literature on Bayesian inference methods in regional science. This discussion is presented in the context of the Spatial Durbin Model (SDM) with heteroskedasticity as a canonical example. The overall performance of different hierarchical models is analyzed. We extend the benchmark specification to the dynamic panel data model with spatial dependence. An empirical illustration of the flexibility of the Bayesian approach is provided through the analysis of the role of knowledge production and spatiotemporal spillover effects using a space-time panel data set covering 49 US states over the period 1994–2005.

The chapter discusses generalized method of moments (GMM) estimation methods for spatial models. Much of the discussion is on GMM estimation of Cliff-Ord-type models where spatial interactions are modeled in terms of spatial lags. The chapter also discusses recent developments on GMM estimation from data processes which are spatially α-mixing.

In regional science, many attributes, either social or natural, can be categorical. For example, choices of travel mode, presidential election outcomes, or quality of life can all be measured (and/or coded) as discrete responses, dependent on various influential factors. Some attributes, although continuous, are subject to truncation or censoring. For example, household income, when reported, tends to be censored, and only boundary values of a range are obtained. Such categorical and censored variables can be analyzed using econometric models that are established based on the concept of “unobserved/latent dependent variable.” The previous examples also share another common feature: when data is collected in a spatial setting, they are all inevitably influenced by spatial effects, either spatial variation or spatial interaction. In contrast to panel data or time-series data, such variation or dependencies are two-dimensional, making it even more complicated. The need for investigating such limited and censored variables in a spatial context compels the quest for rigorous statistical methods.This chapter introduces existing methods that are developed to analyze limited and censored dependent variables while considering the spatial effects. Different model specifications are discussed, with an emphasis on discrete response models and censored data models. Different types of spatial effects and corresponding ways to address them are then discussed. In general, when the spatial variation is of major concern, geographically weighted regression is preferred. When the spatial dependency is the primary interest, spatial filtering and spatial regression should be chosen. Techniques popularly used to estimate spatial limited variable models, including maximum simulated likelihood estimation, composite marginal likelihood estimation, and Bayesian approach, are also introduced and briefly compared.

This chapter provides a survey of the existing literature on spatial panel data models. Both static and dynamic models will be considered. The chapter also demonstrates that spatial econometric models that include lags of the dependent variable and of the independent variables in both space and time provide a useful tool to quantify the magnitude of direct and indirect effects, both in the short term and long term. Direct effects can be used to test the hypothesis as to whether a particular variable has a significant effect on the dependent variable in its own economy and indirect effects to test the hypothesis whether spatial spillovers exist. To illustrate these models and their effects estimates, a demand model for cigarettes is estimated based on panel data from 46 US states over the period 1963–1992.

Spatial interaction or gravity models have been used in regional science to model flows that take many forms, for example, population migration, commodity flows, traffic flows, and knowledge flows, all of which reflect movements between origin and destination regions. This chapter focuses on spatial autoregressive extensions to the conventional least-squares gravity models that relax the assumption of independence between flows. These models, proposed by LeSage and Pace (2008, Spatial econometric modeling of origin-destination flows. J Reg Sci 48(5):941–967, 2009), define spatial dependence in this type of setting to mean that larger observed flows from an origin region A to a destination region Z are accompanied by (i) larger flows from regions nearby the origin A to the destination Z, say regions B and C that are neighbors to region A, which they label origin dependence; (ii) larger flows from the origin region A to regions neighboring the destination region Z, say regions X and Y, which they label destination dependence; and (iii) larger flows from regions that are neighbors to the origin (B and C) to regions that are neighbors to the destination (X and Y), which they label origin-destination dependence. Spatial spillovers in these models can take the form of spillovers to both regions/observations neighboring the origin or destination in the dyadic relationships that characterize origin-destination flows as well as network effects that impact all other regions in the network. We set forth a simulation approach for these models that can be used to produce scalar expressions for the various types of spillover impacts that arise from changes in the explanatory variables of the model.