Thirty-five years of R(S)UE: A retrospective

er.com/locate/regec
Regional Science and Urban Economics 37 (2007) 434–449
www.elsevi

Thirty-five years of R(S)UE: A retrospective☆

Gianmarco I.P. Ottaviano ⁎, G. Alfredo Minerva

University of Bologna, Italy

Accepted 1 January 2007Available online 12 April 2007

Abstract

The aim of this paper is to provide a quantitative retrospective of thirty-five years of Regional Scienceand Urban Economics, from Volume 1 (1971) to Volume 35 (2005). In doing so, the paper follows theintellectual development of the journal through its different editors by comparing their editorial statementswith the actual changes in the characteristics of articles they selected for publication.© 2007 Elsevier B.V. All rights reserved.

JEL classification: R10Keywords: Regional economics; Spatial econometrics; Urban economics

1. Introduction

Thirty-five years may be the appropriate age to evaluate what one has achieved halfway intolife and think about what comes next. At least this was the opinion of Dante when he wrote hisDivine Comedy, a midlife metaphorical account of past experiences and future prospects.Mutatismutandis, this issue of Regional Science and Urban Economics presents a similar, though lesspoetic, effort — one we hope is occurring much earlier than the mid-life of the journal.

While other contributions in this issue look toward the future, the aim of this paper is to take alook back and provide a quantitative retrospective of thirty-five years of Regional Science andUrban Economics (henceforth, RSUE), from Volume 1, dated 1971, to Volume 35, dated 2005.To be precise, the original name of the journal was Regional and Urban Economics, which wasreplaced by its current name in 1975 for editorial reasons we will also discuss.

☆ We thank Richard Arnott and Jacques Thisse for their helpful comments.⁎ Corresponding author. Dipartimento di Scienze Economiche, Piazza Scaravilli 2, 40126 Bologna, Italy. Tel.: +39 051

2098873; fax: +39 051 2098040.E-mail address: [email protected] (G.I.P. Ottaviano).

0166-0462/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.regsciurbeco.2007.01.007

mailto:[email protected]

http://dx.doi.org/10.1016/j.regsciurbeco.2007.01.007

435G.I.P. Ottaviano, G.A. Minerva / Regional Science and Urban Economics 37 (2007) 434–449

The paper follows the intellectual development of the journal through its different editors bycomparing their editorial statements with the actual changes detectable in the articles they selectedfor publication. In particular, Section 2 discusses the editorial turnover. Section 3 presents thecharacteristics of the articles we have chosen to analyze and their classifications. Section 4provides descriptive information on those characteristics and Section 5 summarizes the main factsthat emerge from their analysis. Finally, Section 6 concludes, arguing that the overall intellectualtrajectory of the journal is characterized by a consistent and conscious drive towards improvingthe analytical methods of spatial economics. Such a drive has its roots in the first editorialstatement of RSUE and may explain the frequent perception of RSUE as the most technical of thejournals in the field of regional and urban economics.

2. Editors and editorials

The scientific interests of the editors play a key role in shaping a journal. It is thereforeinteresting to begin with an overview of the editorial boards that have managed RSUE and theireditorial statements at the time they took responsibility.

When the journal was founded, as Regional and Urban Economics, in 1971, the editors wereJean H.P. Paelinck (Netherlands Economic Institute, Rotterdam, Netherlands), Aaron Fleisher(MIT, Cambridge, USA), and Nikolay N. Nekrasov (Council for Studying Productive Forces,Moscow, USSR). The editorial statement (Paelinck, 1971) described the articles in the first issueas ranging “from theoretical–operational over the pure econometric field to the more appliedeconometric exercises.” It suggested that “the various models presented might be useful at onestage or another of regional economic analysis, parameter estimation, and planning” and stressedthe importance the editors attached to improving the analytical methods of spatial economics.

In 1975, Åke Anderson and Walter Isard (both at the Regional Science Department of theUniversity of Pennsylvania in Philadelphia, USA) took over the editorial responsibility. At thattime, the name of the journal changed to Regional Science and Urban Economics. The rationalefor this change can be found in their editorial statement (Anderson and Isard, 1975): “This journalis devoted to the presentation of major advances in Regional Science, a truly interdisciplinaryscience, which has developed very rapidly in the last twenty-five years. The new editors desirethat the interdisciplinary character of Regional Science be fully reflected in this journal throughcontributions by those trained as economists, quantitative geographers, planning analysts, andother regional scientists. The journal is also designed to deepen the foundations of RegionalScience, the creation of new theories and models and their application to forecasting, planningand policy. […] Homogeneity of the journal is also a necessity, which we wish to ensure byconcentrating on contributions that employ formal methods from mathematics, econometrics,operations research and related fields.” Accordingly, the inclusion of the word “science” in thename of the journal served the purpose of accomplishing a transition from Regional Economics toRegional Science, perceived as a richer interdisciplinary field built on a variety of culturalbackgrounds and research techniques.

In keeping with the statement of the founding editors, issues had to be analyzed at a high levelof generality using formal methods of investigation. However, contributions were also required tofocus on immediate or potential uses in terms of forecasting, planning, and policy design at theregional and urban levels.

In 1980, Urs Schweizer (University of Bonn, Federal Republic of Germany) became themanaging editor of RSUE, with Anderson and Isard still serving as co-editors. The editorialpolicy remained substantially unchanged until 1987 when John M. Quigley (University of

436 G.I.P. Ottaviano, G.A. Minerva / Regional Science and Urban Economics 37 (2007) 434–449

California, Berkeley, USA) and Konrad O. Stahl (initially at the University of Dortmund and laterat the University of Mannheim, Germany) took over. Under their guidance, the journal wasrefocused. In their editorial statement (Quigley and Stahl, 1987) they wrote that they wanted “toemphasize the journal's focus on microeconomic analyses of spatial phenomena.” They solicited“original research contributions in spatial economics, economic geography, and relateddisciplines,” mainly dealing with market organization in space, housing and labor markets,transportation, and local public economics. Accordingly, they envisaged a more specializedjournal with a core consisting of microeconomic analyses, and, in their statement, they did notmention the interdisciplinary approach of their predecessors.

The last change in the editorial board took place in 2003 with the appointment of RichardArnott (Boston College, Boston, USA) and Antonio Ciccone (Universitat Pompeu Fabra,Barcelona, Spain). Their statement (Arnott and Ciccone, 2003) acknowledged the editorialreorientation of Quigley and Stahl: “stimulated by issues related to European integration, therehas been a shift from urban to regional topics; advances in computing, applied econometrictheory, and data collection have shifted the balance from theoretical to empirical work; andadvances in economic theory, especially in game theory and growth theory, have been reflected inthe theoretical papers published.” Sharing their predecessors' approach, the new editorsannounced no major changes in the editorial policy, though, as we will see, some novelties beganto appear in the characteristics of the published articles.

To summarize, while all editorial boards stressed the importance of both theoretical andapplied analyses, RSUE underwent a clear transition around the mid-1980s from aninterdisciplinary outlet for advances in Regional Science to a more focused journal with adominant microeconomic content.

3. Variables and classifications

To characterize the evolution of the journal in terms of objective criteria, we present aquantitative investigation of its thirty-five-year history. The investigation is based on a datasetconstructed by analyzing all published articles, from Volume 1, which appeared in May 1971, toVolume 35, released in December 2005. Only research contributions are considered, so weneglect all introductory papers to special issues. The remaining articles are then classified in termsof authors, institutions, JEL topics, disciplines, and methodologies.

3.1. Authors and institutions

The first set of variables refers to authors' characteristics: the number of co-authors, the lengthof papers in terms of published pages, the country of the institution authors are affiliated with, andthe institutions themselves. The most cited papers, as of October 2005, are also identified usingthe ISI Social Sciences Citation Index, which supplies information about the number of citationsper year.

3.2. Topics (JEL)

The second set of variables concerns the subjects of the articles. The most straightforward wayto analyze them is to resort to the JEL Classification System. Unfortunately, RSUE papers reportJEL classification codes only from Volume 23, Number 5 onward. JEL information is thereforesupplemented by the two following additional classifications: “discipline” and “contents.”


3.3. Disciplines and contents

As discussed in Section 2, the journal acted as an interdisciplinary outlet for many years. So,our first additional classification of subjects is designed to capture the disciplines to which thearticles belong. By “discipline” we mean a subject people study, or are taught.1 In this respect, forexample, “econometrics” is a different discipline from “economics.” This leads to the followingsix (mutually exclusive) categories:

1) regional science, input–output models;2) regional and urban economics;3) spatial competition, location theory, monocentric-city models;4) economic geography;5) econometric regional models, spatial econometrics;6) experimental economics.

A few words are necessary to explain these categories. The main criterion we use is to allocate tothe same category papers that are likely to be taught together in courses that are intensive inquantitative methods. The distinction between “regional science” and “regional and urbaneconomics” lies in the microeconomic foundations: while the latter category is based on optimizingagents, the former is not. In particular, “regional and urban economics” includes, among others, allpapers that qualify as R10, R11, R12, R13, R21, and H71, in JEL terms.2 Input–output modelspertain instead to “regional science.” To “spatial competition, location theory, monocentric-citymodels” belong the articles in which distance among locations in space has a geometric formalization(usually in terms of Euclidean distance).3 This may also apply to spaces that are not physical, such asthe space of characteristics in product differentiation theory. Accordingly, some RSUE contributionsfall into the domain of Industrial Organization. Papers in “economic geography” simultaneouslystudy trade in goods and where factors of production locate. Articles in “econometrics” are eitherempirical works, involving simply some statistical estimation, or works presenting new insights ineconometric theory applied to spatial phenomena. The category “experimental economics” is addedfor completeness, even though experiments are very uncommon in regional and urban investigations.

The second additional classification of subjects focuses more on actual contents. For instance,a paper in urban economics might help the reader to understand either local public finance issuesor land use and housing problems. The same might hold for an econometric paper. The nine(mutually exclusive) categories of “contents” are:

1) local public finance, financing of local utilities, tax competition;2) economic integration, migration;

1 This is the definition of the word “discipline” in the Oxford English Dictionary.2 Explanation of the JEL codes is as follows: R10 (General); R11 (Regional Economic Activity: Growth, Development,

and Changes); R12 (Size and Spatial Distributions of Regional Economic Activity; Interregional Trade); R13 (GeneralEquilibrium and Welfare Analysis of Regional Economies); R21 (Housing Demand); H71 (State and Local Taxation,Subsidies, and Revenue).3 The monocentric-city model does belong to urban economics. However, we choose to classify it together with spatial

competition and location theory because its framework has been used extensively in both spatial competition and locationtheory, and it is likely to be taught together with them in courses intensive in quantitative methods. In the classificationsorted by “contents,” we create instead a “spatial competition, spatial pricing and location theory” category, separate fromurban themes.


3) urban economic structure, land use and housing, congestion, segregation;4) regional economic structure, growth, interregional trade;5) spatial competition, spatial pricing, location theory;6) industrial districts, technological externalities, innovation studies;7) transportation, logistics, infrastructure (their planning included);8) regional and urban planning, regional policy, regulated local markets;9) environmental economics.

Note that the journal does not publish articles that analyze specific policies, except to theextent that the article makes a methodological contribution.

3.4. Methodologies

The final variable deals with methodological issues. Depending on their research methods,articles are classified into the following five categories:

1) analytical;2) verbal;3) applied;4) theoretical econometrics;5) computational methods and simulation.

These categories are not mutually exclusive, so combinations of them are possible.“Analytical” articles are theoretical analyses that consider problems at a high level of generalityusing mathematical tools without relying on numerical methods. In “verbal” articles, topics arediscussed without resorting to mathematical or statistical methods. For instance, a review of therelevant literature on a particular issue is classified as “verbal.” In “applied” articles, a standardframework (from econometrics, linear programming, etc.) is implemented on real (notautomatically-generated) data, with the purpose of describing the features of a particular set ofindividuals (regions, countries, markets, etc.). Papers are classified under “theoreticaleconometrics” when they present novel results concerning estimation procedures or tests.4

Even though these papers often entail some application to real data, they are not labeled as“applied,” as long as the application is made mainly for illustrative purposes. Finally, the category“computational methods and simulation” contains articles whose results are obtained throughcomputational programs (e.g., simulations of econometric models, Monte Carlo methods,solutions of algorithms).

Theoretical articles differ widely in terms of the mathematics they use. A sub-classification istherefore proposed, with the aim of mapping the specific types of formal tools they employ.Specifically, once identified as “analytical,” an article is subsequently classified into one of 13categories (see Table 1).

Some comments are in order. First, our classification combines purely mathematical categorieswith others drawn from mathematical economics. Second, categories are not necessarily mutuallyexclusive. For instance, both models with “asymmetric information and incomplete contracts”and “voting models” belong to “game theory.” However, in many cases it is interesting to know

4 The theoretical econometrics category includes not only spatial econometrics papers, but, also, the (few) papersdealing with other branches of econometrics. More information on this issue is presented in Section 4.

Table 1Classification of analytical tools

General equilibriumExpectations based general equilibriumPartial equilibriumGame theoryAsymmetric information and incomplete contractsVoting modelsProbability theory; stochastic processes; discrete choice modelsSet theory and topologyLinear algebra and matrix theoryGraphs; algorithms; networksOrdinary and partial differential equations; difference equationsOperations research and mathematical programmingOptimal control and calculus of variations


the type of game involved. Third, general equilibrium papers with some treatment of expectationsare singled out. Fourth, optimization methods are a building block of economic theory, so theyappear in virtually all general and partial equilibrium models. The labels “general equilibrium” or“partial equilibrium” are attached to articles featuring optimization procedures that arestraightforward (e.g., they involve a standard Lagrangean). When the optimization procedure isinstead more sophisticated, articles are classified as “mathematical programming,” if the programis static, and “optimal control,” if the program is dynamic (e.g., it involves a Hamiltonian).

4. Descriptive statistics and trends

We are now ready to present the main facts and trends that emerge from our database. In doingso, variables are discussed in the same order as they were introduced in the previous section.

4.1. Authors and institutions

Most authors are affiliated with institutions belonging to North America. Together, the U.S.and Canada account for 660 out of 1062 articles. Europe accounts for 338 papers, with theNetherlands ranking at the top, with 56 papers. Japanese affiliations are attached to 67 articles. Allremaining countries comprise 88 papers in total. For articles with more than one author, since co-authors may come from different parts of the world, we count the affiliation of each co-authorindependently. Hence, the sum of the above figures exceeds the total number of published papers.

Fig. 1 plots the evolution of authors' geographical affiliations through time.5 To smooth data, athree-year moving average is computed. Initially, European authors were predominant (with theircore located in the Netherlands). North-American authors took the lead in 1975 and kept it fortwenty-five years until 2000 when the numbers of North-American and European contributorsconverged. It is interesting to note that the opening and the widening of the gap between NorthAmerica and Europe coincided with the first two significant editorial changes in 1975 and 1987,respectively. The closing of the gap instead predates the third and last editorial turnover in 2003.

As expected from the above pattern, the most active institutions belong to the U.S. Amongthem, the most represented is the University of Pennsylvania (46 papers), followed by the

5 On their vertical axes, all figures report the shares of papers with a certain characteristic per volume.

Fig. 1. Geographical affiliation of authors (three-year moving average).


University of Illinois and Ohio State University (41 and 26 papers, respectively). Then, there aretwo European institutions, the Free University of Amsterdam (24) and the Catholic University ofLouvain (23). Only two other institutions feature at least 20 papers: Northwestern University (21)and the University of Tokyo (20).

Over the entire thirty-five-year period, the average number of authors per paper equals 1.6.However, plotting the number of authors per paper against the volume number reveals a strong,upward time-trend. This trend matches the well-known, stylized fact that the number of authorsper published paper in economics has been increasing over time. For instance, in Volume 5 theaverage number of authors was 1.3, while in Volume 35 it had risen to 2, the highest value ever.

The trend regarding the length of articles, in terms of average published pages, is, instead, non-monotonic. Of course, the length of papers also depends on typographical arrangements that havechanged over time. Nonetheless, provided that such changes have not been dramatic, there seemsto be a downward trend in published pages until Volume 15 (15.7 pages) followed by an upwardtrend until Volume 26, after which their number becomes stable at around 22 pages. The totalnumber of pages per volume increased twice, as the outcome of deliberate editorial choices: from400 to 600 pages starting with Volume 10 (1980); from 600 to 800 pages starting with Volume 23(1993).

A small subset of papers had a larger impact than others on the profession. This appears rathermarkedly when one looks at the total number of citations as of October 2005. Clearly, thisunderestimates the potential impact of recent papers and does not adjust for the “quality” ofcitations.6 The articles that received more than 50 citations as of October 2005 are listed in Table 2.The most cited paper is Fujita and Ogawa's “Multiple equilibria and structural transition of non-monocentric urban configurations” (Volume 12, 1982). Two of the most cited papers (the first byBucovetsky andWilson, 1991, the second byWilson alone, 1991) belong to a special issue on localpublic economics (Volume 21, edited by David Wildasin and John Wilson).

Looking at the evolution of citation records by year gives us information about the diffusion ofthe articles and the popularity of their topics. Since a systematic investigation of these phenomena

6 Considering the unadjusted total number of citations per article has several shortcomings. Their discussion is,however, beyond the scope of the present paper.

Table 2Citation records of most cited papers

Rank Citations Article Authors

1 106 Multiple equilibria and structural transition of non-monocentric urbanconfigurations, 12, 161–196.

Fujita M., Ogawa H.

2 100 The baby boom, the baby bust, and the housing market, 19, 235–258. Mankiw N. G., WeilD. N.

3 94 Minimum information principle — theory and practice, 7, 137–168. Snickars F., WeibullJ. W.

4 85 Subcenters in the Los Angeles region, 21, 163–182. Giuliano G., SmallK. A.

5 83 Tax competition with two tax instruments, 21, 333–350. Bucovetsky S.,Wilson J. D.

6 78 The contribution of publicly provided inputs to states economies, 22, 229–241. Garcia-Milà T.,McGuire T. J.

7 67 A new look at static spatial price equilibrium models, 12, 579–597. Florian M., Los M.8 64 Estimating the demand for housing characteristics — a survey and critique, 15,

77–107.Follain J. R.,Jimenez E.

9 64 Location and theory of production — review, summary and critique of recentcontributions, 8, 117–128.

Miller S. M., JensenO. W.

10 63 Simple diagnostic tests for spatial dependence, 26, 77–104. Anselin L., et al.11 59 Tax competition with interregional differences in factor endowments, 21, 423–

451.Wilson J. D.

12 56 Increasing returns, monopolistic competition, and agglomeration economies inconsumption and production, 18, 125–153.

Rivera-Batiz F.


is beyond the scope (and the room) of the present retrospective, Fig. 2 presents the diffusionpatterns of the four most-cited papers only (a three-year moving average is computed). Three outof the four papers (Mankiw and Weil, 1989; Snickars and Weibull, 1977; Giuliano and Small,1991) exhibit a single peak.7 The pattern of diffusion is such that citations reach their peak someyears after publication and then fade away with time. Only the paper by Fujita and Ogawa (1982)features three peaks (the third is still growing), which shows that the paper has been repeatedly“rediscovered.” Indeed, we may say it is an “evergreen”! Considering the raw (not averaged) data,its three peaks appear in 1988, 1996, and 2004.

4.2. Topics (JEL)

As mentioned earlier, JEL codes are available only from Volume 23 Number 5, i.e., for a totalof 406 papers, less than half of the total number of observations. So, JEL codes provideinformation only on relatively recent patterns. Table 3 reports the frequency of JEL categoriesmentioned in at least 20 papers.

The first two categories (R12 and R11) are “general purpose” and, thus, are not veryinformative in their own right. Nevertheless, they identify the two pillars of the journal: locationtheory and regional development (see also R30 and R10 down in the Table). The third category(L13) reveals the considerable attention paid to oligopoly and imperfect competition, typicalsubjects of Industrial Organization. This category captures most of the contributions related tospatial competition. The fourth category (R13) is also a rather comprehensive cell that signals the

7 In the case of Snickars and Weibull (1977), the lack of citation data from 1996 to 2001 generates an additional smallpeak in recent years.

Fig. 2. Diffusion patterns for most cited papers (three-year moving average).


relevance of endogenous factor price determination in a regional context. The fifth and sixthcategories (R21 and R14) are typical of the fact that urban economics is concerned with the landmarket, as well. Another category that is well represented is trade models with imperfectcompetition (F12). This is interesting for two reasons. On the one hand, it confirms theimportance of imperfect competition in the journal's output (see L13 above). On the other hand, ithighlights the interest that the journal has for International Trade. Finally, among the mostfrequent JEL categories, we find topics in Labor Economics (R23) and Public Economics (H71),mostly related to local public finance.

4.3. Disciplines and contents

Given the partial coverage of the JEL classification, it is useful to look at the two additionalclassifications of subjects we proposed in Section 3.3. Their advantage is that they can be appliedto the full set of articles.

Table 4 reports the frequency of papers by “discipline” (i.e., by a subject people study, or aretaught). Overall, RSUE is predominantly a “regional and urban economics” journal: 44% ofpapers (468 out of 1062) belong to this category, which matches the picture that emerged from the

Table 3Ranking of the most frequent JEL categories

Rank Frequency JEL Code Description

1 51 R12 Size and spatial distributions of regional economic activity; Interregional trade2 38 R11 Regional economic activity: growth, development, and changes3 37 L13 Oligopoly and other imperfect markets4 32 R13 General equilibrium and welfare economic analysis of regional economies5 26 R21 Housing demand6 25 R14 Land use patterns7 22 F12 Models of trade with imperfect competition and scale economies7 22 R30 Production analysis and firm location — general9 20 R10 General regional economics — general9 20 R23 Regional migration; regional labor markets; population9 20 H71 State and local taxation, subsidies, and revenue

Table 4Ranking of disciplines

Rank Frequency Description

1 468 Regional and urban economics2 208 Econometric regional models; spatial econometrics3 165 Regional science; input–output models4 159 Spatial competition, location theory, monocentric-city models5 23 Economic geography6 1 Experimental economics– 38 Papers not classified


JEL codes. Econometric analyses are also important, covering 20% of the papers. The other tworelevant disciplines are “regional science, input–output models” and “spatial competition,location theory, monocentric-city models,” each accounting for 15% of the published papers.Since the former category is empirically oriented while the latter category is theoretical, the totalweight of theoretical analyses is roughly 60% of the output (i.e., 44% of “regional and urbaneconomics” plus 15% of “spatial competition, location theory, monocentric-city models”).Empirical analyses (“regional science, input–output models” and “econometrics”) amount to 35%of the total articles.

Fig. 3 depicts the evolution of the disciplines through time (plotted data are smoothed by athree-year moving average). At the outset of the journal in 1971, more than 70% of the articlesbelonged to “regional science”with “regional and urban economics” covering another 20%. Sincethen, the share of the former, generally, has been falling over time, with the latter taking the leadfor the first time in 1974. The exception is the editorial period of Anderson and Isard, which beganin 1975. As previously discussed in Section 2, these editors introduced the word “science” into thename of the journal and stressed the specificity of Regional Science as “a truly interdisciplinaryfield.” The result of their efforts is clearly visible in Fig. 3, which shows “regional science”increasing again and “regional and urban economics” falling back until 1980, the year Schweizerbecame managing editor. Since then, the number of papers in regional science and input–outputanalyses has been declining steadily and by the 1990s their importance had already becomeconsiderably limited.

The opposite trend characterizes “econometrics.”8 While during the early-1970s very feweconometric papers appeared in the journal, their number increased steadily during the 1980s andthe 1990s. Indeed, in the last years, especially under Arnott and Ciccone, econometriccontributions have essentially caught up with those in regional and urban economics. Finally,most likely due to its mixed nature, the category “spatial competition, location theory,monocentric-city models” does not seem to show any particular trend.

Turning to the classification in terms of “contents,” the corresponding frequencies are shown inTable 5. The most important category concerns urban issues, with a share of 27% (291 out of 1062papers). The second and third ranks are occupied by regional analyses (19%) and by contributionsin local public economics (12%), respectively. The remaining distribution is more even,suggesting a rather diversified portfolio of subjects (spatial competition, spatial pricing andlocation theory; transportation; planning; economic integration and migration; industrial clusters;environmental issues). Lastly, the high number of papers not classified largely consists of articleswhose main contributions concern technical and methodological issues.

8 Remember that this category contains both applied and theoretical contributions.

Fig. 3. Time trends of top disciplines (three-year moving average).


Fig. 4 describes the evolution of the four main contents over time. In 1971, papers analyzingregional issues were dominant, and accounted for 45% of the total. The second category, dealingwith urban issues, covered a share of 15% only. However, by the end of the 1970s the gapbetween the two categories had already vanished, and by the mid-1980s urban topics took thelead. This trend reached its climax in the second half of the 1990s, with urban topics at 35% andregional topics at 10%. Two factors seem to drive this evolution. First, by the end of the 1980s andthe beginning of the 1990s two of the most cited papers (Mankiw and Weil, 1989; Giuliano andSmall, 1991), both dealing with urban issues, appeared in press and gave rise to a vigorous debate.For instance, Volume 21 hosts an issue almost exclusively composed of papers questioning thefindings of Mankiw and Weil. Second, many special issues on urban structure and housing beganto be edited.9 As these two drivers lost their strength, the last few years have witnessed a renewedconvergence in the numbers of regional and urban papers.

While spatial competition again exhibits a substantial stability during the last two decades, apositive trend characterizes the evolution of local public finance, whose relevance, through peaksand troughs, has been consistently rising. Two peaks can be identified, in correspondence withtwo special issues. The first is the aforementioned special issue edited by Wildasin and Wilson in1991 (Volume 21). The second is a special issue edited by Konrad Stahl in 2001 (Volume 31) on“Fiscal Competition and Federalism in Europe.” Nevertheless, in recent years, the relevance oflocal public finance has gone beyond those special issues, as testified, for instance, by the densityof local public finance papers in the last volume analyzed (Volume 35). In particular, fiscalcompetition theory (dealing with the horizontal and vertical strategic interactions amonggovernments) is one of the journal's leading niches. Its importance has been growing in the wakeof the debates in the European Union and shows that, by now, the journal has established itself asa respected outlet in local public finance.

9 From 1990 to 1999, out of twelve special issues, five were dedicated to urban themes.

Table 5Ranking of contents


1 291 Urban economic structure; land use and housing; congestion; segregation2 204 Regional economic structure; growth; interregional trade3 132 Local public finance; financing of local utilities; tax competition4 77 Spatial competition; spatial pricing; location theory5 72 Transportation; logistics; infrastructure (their planning included)6 66 Regional and urban planning; regional policy; regulated local markets7 41 Economic integration; migration8 25 Industrial districts; technological externalities; innovation studies9 16 Environmental economics– 138 Papers not classified


4.4. Methodologies

Section 3.4 proposed a classification of papers according to five research methods (i.e.,“analytical,” “verbal,” “applied,” “theoretical econometrics,” and “computational methods andsimulation”). The corresponding frequencies are listed in Table 6.10 The table shows thatabstract mathematical modeling is the most frequent method, since the “analytical” categorycontains 66% of the papers (700 out of 1062). “Applied” papers, using real data, instead, cover38% of the total (397 papers). Some articles contain both a theoretical side and an application toreal data. Their exact number is 99, almost 10% of the total. The share of the remainingmethodologies (“computational methods,” “verbal,” and “theoretical econometrics”) is limitedto 14%.

Theoretical econometrics deserves specific attention. As documented elsewhere in this specialissue (see Anselin's article), for many years, RSUE was almost the only economics journal toconsistently publish papers in spatial econometrics. Among 36 papers with new theoreticaleconometrics insights, 22 can be identified as purely theoretical. Almost all of them (19) belong tospatial econometric theory.

The evolution of the relative importance of the different methods over time is depicted inFig. 5. The presence of analytical and applied papers is roughly balanced both at the beginningand at the end of the observed period. However, in the meantime, the relative number of appliedpapers exhibits a non-monotonic behavior, first falling during the 1970s and then catching upagain during the last decade when the number of papers with both theoretical and empiricalapplications increases also. In contrast, computational methods and simulation techniques play amarginal role and exhibit no noticeable trend.

Analytical papers are further classified into sub-categories, depending on the specificanalytical tools they use. The corresponding frequencies are reported in Table 7. “Generalequilibrium” is the dominant tool, which reflects the centrality of endogenous factors' prices inregional models. The second and third categories, “partial equilibrium” and “game theory,” coverthe main tools of spatial competition models. A perception that seems to have persistedthroughout the journal's life identifies RSUE as one of the most technical outlets in regional

10 Categories are not mutually exclusive, so some papers are allocated to more than one cell. This explains why the sumof papers across the categories exceeds the number of published papers.

Fig. 4. Time trends of top contents (three-year moving average).


science and urban economics. This is corroborated by the wide array of analytical tools ranked inTable 7.

5. Overview of main facts

Overall, six facts stand out in the history of the journal. First, after its debut, a quick transitiontook place from regional science to urban economics. Regional science papers flourished duringthe early years of Anderson and Isard. On the contrary, in the last twenty years, the journal hasbeen mainly concerned with regional and urban economics, as well as econometrics. This hasweakened the interdisciplinary spirit of the early years. The turning point can be located at thebeginning of the 1980s, and the transition was virtually accomplished by the end of the decade,under the impulse of Quigley and Stahl.

Second, the importance of econometric contributions (whether theoretical or applied) hasincreased steadily through time. Today, under Arnott and Ciccone, their relevance is only slightlyless than that of purely theoretical analyses. A key niche is spatial econometrics. Indeed, for manyyears RSUE was almost the only economics journal to consistently publish spatial econometricpapers. An important recent trend, which mirrors what is happening elsewhere in economics, isthe growing use of increasingly disaggregated data. A few papers in this special issue highlightthe fact that the availability of detailed data sets at the micro-geographic level is boosting newresearch in regional economics.

Table 6Ranking of research methods


1 700 Analytical2 397 Applied3 63 Computational methods and simulation4 48 Verbal5 36 Theoretical econometrics

Fig. 5. Time trends of top methodologies (three-year moving average).


Third, fiscal competition theory is another key niche of the journal. While its importance is stillgrowing in the wake of the debates in the European Union, by now the journal has establisheditself as a respected outlet in local public finance.

Fourth, while authors from European institutions dominated the early years of the journal,RSUE's history has been mainly written by contributors with North-American affiliations. Onlyrecently has the trans-Atlantic gap vanished. This is possibly due to the centrality that regionalissues have gained in the process of creating, deepening and enlarging the European Union, thusattracting the efforts of researchers (especially younger ones) all over Europe.

Finally, two facts are particularly puzzling: the relative paucity of papers using computationalmethods and simulation, and the scarcity of contributions belonging to the so-called “neweconomic geography.”

As in other domains of economics, and also in regional economic research, some scholars(e.g., Nijkamp in this special issue) have questioned the validity of the deductive and equilibrium

Table 7Ranking of analytical tools


1 236 General equilibrium2 116 Partial equilibrium3 113 Game theory4 66 Operations research and mathematical programming5 53 Probability theory; stochastic processes; discrete choice models6 22 Linear algebra and matrix theory7 21 Optimal control and calculus of variations8 18 Ordinary and partial differential equations; difference equations9 15 Graphs; algorithms; networks10 14 Asymmetric information and incomplete contracts11 12 Voting models12 6 Expectations-based general equilibrium13 3 Set theory and topology– 367 Papers not classified


approach as the appropriate methodology to be used to analyze the space-economy. In their view,the entire subject should be reconsidered in light of recent advances in complexity theory.

Looking at our database, evolutionary algorithms and computer simulations still have a longway to go in the journal. Computational methods and simulations account for only a tiny share ofthe total papers, and there is no sign of any upward trend. In Table 7, only 15 papers employgraphs, algorithms, or networks. In this respect, it is remarkable that, in a journal that claims to beexclusively concerned with spatial economic phenomena, practically none of the papers employmaps. This trend will probably change, now that mapping software is becoming moresophisticated and easier to use.

“New economic geography” (henceforth, NEG) is a line of research that applies the generalequilibrium framework of international trade models to spatial issues, with a special focus onimperfect market structures, usually interpreted in terms of monopolistic competition. Itshallmark is endogenous agglomeration driven by pecuniary externalities whose strength dependson the level of various microeconomic parameters such as the extent of product differentiation orthe obstacles to the geographical mobility of goods and factors. By now, several surveys of NEGare available. Two advanced textbooks (Fujita et al., 1999; Baldwin et al., 2003) systematizeNEG's main results, thus signaling its achieved maturity. The reference list of Baldwin et al.(2003) can be used to assess the impact of NEG on RSUE, and vice versa. Since the birth of NEGtraditionally coincides with the publication of Krugman (1991), we focus only on references inscientific journals from 1991 onwards. There are 136 such citations, and only nine of them belongto RSUE. Other journals in the field of regional and urban economics do not perform any better,with 12 references in total. This may reveal what some scholars identify as the main contributionof NEG to regional and urban economics: the diffusion of the interest in spatial issues to a moregeneral audience.

Since 2003, under Arnott and Ciccone, there has been an up-tick, but NEG is still far frombeing an important niche for the journal. As argued elsewhere in this special issue (e.g., Behrensand Thisse; Quigley), NEG themes may represent promising waves of the future. In this respect, itis somewhat paradoxical for RSUE that, so far, most NEG papers on local public finance and taxcompetition have appeared in international economics and public economics journals.

6. Conclusion

Over the past thirty-five years, RSUE has published articles characterized by a rich variety ofmethodologies, contents, and authors' geographical origins. Its contributions have ranged fromregional science to spatial econometrics and urban economics, with distinctive niches in spatialcompetition and local, public finance. In recent years, the journal has become more focused onregional and urban economics than on regional science, with an increasing number of articlesfeaturing empirical applications. This focus has somewhat reduced the journal's methodologicaldiversity.

All in all, the intellectual trajectory of RSUE is characterized by a consistent and consciousdrive towards improving the analytical methods of spatial economics. Such a drive has itsroots in the journal's first editorial statement, issued in 1971, and may explain the sharedperception of RSUE as the most technical of the journals in regional and urban economics. Asthis cannot be stated definitely without data on other journals, an interesting direction forfuture work on the intellectual history of the field could be to contrast RSUE with otherjournals, and also to ask urban and regional economists in what ways they view RSUE as beingunique.


References

Anderson, A.E., Isard, W., 1975. Regional Science and Urban Economics — Editorial. Regional Science and UrbanEconomics 5, 3.

Anselin, L., in press. Spatial Econometrics in RSUE: Retrospect and Prospect. Regional Science and Urban Economics 37.Anselin, L., Bera, A.K., Florax, R., Yoon, M.J., 1996. Simple diagnostic tests for spatial dependence. Regional Science

and Urban Economics 26, 77–104.Arnott, R., Ciccone, A., 2003. RSUE Editorial Transition Statement. Regional Science and Urban Economics 33, 651.Baldwin, R., Forslid, R., Martin, P., Ottaviano, G., Robert-Nicoud, F., 2003. Economic Geography and Public Policy.

Princeton University Press, Princeton.Behrens, K., Thisse, J.-F., in press. Regional economics: a new economic geography perspective. Regional Science and

Urban Economics 37.Bucovetsky, S., Wilson, J.D., 1991. Tax competition with two tax instruments. Regional Science and Urban Economics 21,

333–350.Florian, M., Los, M., 1982. A new look at static spatial price equilibrium models. Regional Science and Urban Economics

12, 579–597.Follain, J.R., Jimenez, E., 1985. Estimating the demand for housing characteristics — a survey and critique. Regional

Science and Urban Economics 15, 77–107.Fujita, M., Ogawa, H., 1982. Multiple equilibria and structural transition of non-monocentric urban configurations.

Regional Science and Urban Economics 12, 161–196.Fujita, M., Krugman, P., Venables, A., 1999. The Spatial Economy: Cities, Regions and International Trade. MIT Press,

Cambridge, MA.Garcia-Milà, T., McGuire, T.J., 1992. The contribution of publicly provided inputs to states economies. Regional Science

and Urban Economics 22, 229–241.Giuliano, G., Small, K.A., 1991. Subcenters in the Los Angeles region. Regional Science and Urban Economics 21, 163–182.Krugman, P., 1991. Increasing returns and economic geography. Journal of Political Economy 99, 483–499.Mankiw, N.G., Weil, D.N., 1989. The baby boom, the baby bust, and the housing market. Regional Science and Urban

Economics 19, 235–258.Miller, S.M., Jensen, O.W., 1978. Location and theory of production — review, summary and critique of recent

contributions. Regional Science and Urban Economics 8, 117–128.Nijkamp, P., in press. Ceteris paribus, spatial complexity and spatial equilibrium— an interpretative perspective. Regional

Science and Urban Economics 37.Paelinck, J.H.P., 1971. Editorial. Regional Science and Urban Economics 1, 1–2.Quigley, J.M., in press. RSUE: Reflections after 35 years. Regional Science and Urban Economics 37.Quigley, J.M., Stahl, K.O., 1987. Editorial statement by the new editors. Regional Science and Urban Economics 17, 177.Rivera-Batiz, F., 1988. Increasing returns, monopolistic competition, and agglomeration economies in consumption and

production. Regional Science and Urban Economics 18, 125–153.Snickars, F., Weibull, J.W., 1977. Minimum information principle — theory and practice. Regional Science and Urban

Economics 7, 137–168.Wilson, J.D., 1991. Tax competition with interregional differences in factor endowments. Regional Science and Urban

Economics 21, 423–451.

Documents

Thirty-five years of R(S)UE: A retrospective