The New Urban Economics

And Alternatives

H.W. Richardson

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781135683115
• Edition: 1
• Topic: Architecture
• Subtopic: Urban Planning & Landscaping

This book was first published in 1977. Urban economics is a relatively young field of economics; hardly existing except perhaps in real estate and land economics curricula-before the 1960s. Within the last few years, especially after 1 971, there has been a growth of interest in urban economic theory, strong enough even to attract the attention of general economic theorists. These new theoretical writings have been named the 'New Urban Economics'-NUE for short. The aim of this monograph is to survey and assess NUE, to evaluate its contribution to urban economics, to offer a few extensions and to say something about the future direction of the subfield.

1
What is 'New Urban Economics'?

This book focusses on a recent theoretical branch of urban economics, both directly and indirectly (via comparative evaluation of the major alternative approaches to this theory). Mills and MacKinnon (1973) have dubbed this subfield "the new urban economics". As a space saver, this book will use the acronym NUE. This introduction will attempt to distinguish NUE from the rest of urban economics⁽¹⁾.

The boundaries of NUE are very blurred. Some define it very broadly as referring to almost any systematic theorizing about urban spatial structures. Here it is defined rather narrowly as urban economic theories based upon deriving general equilibrium from the principle of utility maximization in a one-dimensional city. According to this view, the partial-equilibrium, utility-maximization models of the early 1960s (Alonso, 1964; Muth, 1961a; 1969) are antecedents of NUE rather than components. Similarly, the more recent work on discrete models (that is, disaggregated two- or three-dimensional cities) by Mills (1972b), Hartwick and Hartwick (1974), MacKinnon (1974), and others falls outside the defined scope⁽²⁾. Since the authors of these papers have also written within the narrower field, the delimitation may be a little confusing. However, the boundary is not arbitrary. The linear-programming models characteristic of the discrete-approach school imply a very different philosophical outlook from the utility-maximization models, with less reliance on market processes and much closer links with the urban planning models of the 1960s—links that have implications for the probability and form of empirical testing.

The narrow definition of NUE is not prompted by the desire to set up an internecine fight between NUE and non-NUE models, but is intended as a label for the distinctiveness of its approach. In effect, NUE represents an attempt by economic theorists to explore the usefulness of the methodology and concepts of mainstream economic theory in the analysis of urban problems. In a sense, as several observers and critics have noticed (Anas and Dendronis, Mills and MacKinnon, and Richardson, 1973a), it is closely analogous to aggregate growth theory, the main difference being that its focus is economic behaviour over space rather than through time. However, in order to contain this analysis within manageable proportions, NUE theorists, like the growth theorists before them, use an artificial vehicle. The 'city' of NUE, like the 'aggregate economy' of growth theory, is much closer to constructs of the imagination than to simplified representations of reality.

To place NUE in historical perspective, a thumbnail sketch of urban economics may be helpful. The intellectual origins of modern urban economics can be traced to a model used for an entirely different purpose—the agricultural location theory of von Thünen (see Hall, 1966). von Thünen's rings for different agricultural crops, where the highest-value crops (in terms of land intensity) are produced closer to the market and there is an inverse relationship between land rent and transport costs, were easily translatable into the concentric zones of urban-rent theory. The description 'von Thünen model' is still used as an approximate term for the standard urban-rent model.

For the first sixty years of this century, most of the writing about urban problems was by either land and real-estate economists (Hurd, Haig, Hoyt, Ratcliff, and others) or by human ecologists and sociologists (Park, Burgess, McKenzie). By inductive reasoning Hurd (1903) stressed the crucial importance of accessibility as a determinant of land values, whereas Haig (1926) highlighted the substitutability between rent and transport costs. Hoyt (1939) provided a link between the land economists and the sociologists by showing the importance of socioeconomic status groups as an influence on urban residential patterns, whereas the Chicago human ecologists demonstrated that social stratification provided an alternative (or, perhaps more accurately, complementary) explanation of residential segregation to income differentials. Although the urban land economists such as Ratcliff (1949) dealt with a whole range of real-estate valuation problems, their work stressed the critical significance of land rent as the instrument for allocating resources within the urban economy. On a different theme, Clark's work (1951) on urban population densities rediscovered the concept of the 'density gradient' that has become so important in urban economic analysis.

At the risk of neglecting other landmarks, it is a fair generalization to say that urban economics 'took off' in the 1960s. One paper prior to 1960 (Beckmann, 1957) pointed the way to the much later NUE models by obtaining a market equilibrium for rents and densities in which rich households live on the urban periphery, although the assumptions used were much more restrictive than are found in more recent models (rent plus commuting costs are a specific function of income, households maximize site size for a given rent expenditure, etc). In a monograph devoted to the measurement of land-value benefits due to transportation improvements, Mohring and Harwitz (1962) employed the simplest form of urban model by assuming Central Business District (CBD) employment and a residential ring in which identical households live on the same lot size. This implies a horizontal density function, a linearly declining rent function, and (as an equilibrium condition) the complementarity of rent and transport costs.

However, the real origins of modern urban economics were the more or less simultaneous appearance of the first versions of studies by Wingo (1961a), Alonso (1960; 1964) and Muth (1961a; 1969). Since their work is reviewed in chapter 2, only the briefest comments are made here. Wingo developed a market-clearing model of the land market based on simple assumptions about household expenditure, namely that rent plus commuting costs equal a constant sum. This reliance on the older Haig assumption makes Wingo's analysis much less a direct antecedent of NUE than either Alonso's or Muth's. Nevertheless, Wingo made important contributions to the analysis of the influence of transportation technology on urban land use, the significance of which has tended to be neglected. Alonso's work was seminal because it extended urban rent theory by developing a new concept—the bid-rent function—and, perhaps even more important in the context of being a forerunner of NUE, by basing the analysis on the utility function of the individual household (and, since his model also applied to nonresidential land use, on the profit function of the individual firm). Alonso's theory reconciled the concentric-zone model of urban land rent with the utility-maximization behaviour of households (and profit maximization by firms). It only failed to go the whole way because it used a partial rather than a general equilibrium approach.

From the same stable as Alonso (the University of Pennsylvania), Herbert and Stevens (1960) developed a linear-programming model that simulated the residential land market by using an optimizing equivalent of the bid-rent function (that is, maximizing rent-paying ability). This was an important step in that it has been taken up in more recent models as providing a viable, and despite difficulties a more operational, approach to the utility-maximizing model. Muth also used a partial-equilibrium model, but the scope of his analysis was very wide, embracing analysis of population densities and the since neglected question of housing supply. He also discussed many of the complications that follow once the simple assumptions of the standard model are dropped. These include heterogeneous preferences, locally employed workers, and multicentric cities, though his analysis was a passing commentary rather than a development of comprehensive solutions. Muth also subjected his propositions to empirical testing (by using a Chicago data base), and attempted to explore their policy implications.

There was then a hiatus of several years in the development of the theory. Urban economists became interested in other themes, especially policy questions. Whether this was a return to reality or a sidetrack is a question of judgment. In any event, the reason for the diversion was quite clear. In the 1960s governments throughout the world, but especially in the united States, became preoccupied with a wide range of urban problems that were summed up in the phrase 'urban crisis'. These preoccupations resulted in a spate of public policy programmes as an attempt to solve some of these problems. Urban economists responded to public interest in these issues, partly for altruistic reasons—to try to help, partly because topicality provides a degree of self-justification, and partly no doubt to take advantage of the boom in urban research contracts. Although the problems analyzed by the policy-oriented urban economists— such as ghetto strategies, the flight from the central city, metropolitan fiscal problems, urban transportation, poverty and housing, crime, metropolitan consolidation, city size control, and national urbanization policy—have theoretical implications, the theoretical aspects were not seriously discussed. However, the concern with policy issues dissipated towards the end of the decade as the problems were discovered to be more intractable than had been expected, as the research and programme funds dried up, and as governments moved on to struggle with 'crises' of a different kind. Many urban economists still retain a strong interest in policy problems, but most would admit that the steam has gone out of these issues.

Returning to the evolution of NUE, a paper by Strotz (1965) was the first to examine how congestion taxes might be used to convert a competitive spatial equilibrium into an optimum within the framework of a utility-maximization model. But the analysis had little impact at the time and languished in obscurity until the early 1970s. The two papers that began NUE in earnest were by Mills (1967) and Beckmann (1969). Mills examined the role of the production, housing, and transportation sectors within the framework of a general equilibrium model for a city. Although the model was too ambitious for the time and could not be solved properly, it raised many important issues that have been recurrent topics of debate in NUE, such as the role of increasing returns, the conditions of locational equilibrium, the allocation of land to transportation, the nature of traffic congestion, the determinants of the urban boundary, the shape of the rent and density gradients, as well as neglected questions such as the importance of capital-land substitution in urban development. Nevertheless Beckmann's model was the trigger that really started NUE. One factor was that the paper appeared in the first volume of a new journal, the Journal of Economic Theory, which has since become one of the most influential vehicles for research in economic theory. Several important NUE papers have appeared in this journal, and the publication of Beckmann's paper there gave the field a respectability among economic theorists that it might otherwise lack. Another reason for the impact of the Beckmann paper was its elegance and simplicity. Although Beckmann was incorrect in certain details, for example, incorrect specification of the boundary conditions (Delson, 1970) and that it is only a special case rather than a general model (Montesano, 1972)⁽³⁾, his model derived rent and density functions for a log-linear utility function, a linear transportation cost-distance function, and a population described by a Pareto income distribution (but having identical utility functions that included distance as an argument).

By 1970 NUE was on the road. There were several important papers in 1970 and 1971, and the field had developed to the extent that symposia were published in 1972 (in the Swedish Journal of Economics) and in 1973 (in the Bell Journal of Economics and Management Science). NUE has attracted the attention not only of distinguished spatial economists, such as Beckmann and Mills, but of leading general economic theorists such as Solow and Mirrlees. This book will attempt to evaluate these and other contributions to this fascinating branch of urban economics.

⁽¹⁾ There are now two review papers in existence, one by Mills and MacKinnon (1973), the other by Anas and Dendrinos (1976).

⁽²⁾ Other observers cut the urban economics cake differently. "The hallmark of the new urban economics is the use of fairly sophisticated mathematics—calculus of variations, programming and control theory—to characterize some fundamental aspects of urban structure ... played by almost entirely new players ... general economic theorists who have recently turned their attention to the urban economy" (Mills and MacKinnon, 1973, page 594).

⁽³⁾ Montesano showed that with a zero transport rate the Beckmann model yields the absurd result that residential density and land rent are increasing functions of distance tending to infinity. Also, income decreases with distance only when income is infinitely large at the edge of the CBD. However, Montesano's restatement salvages the model.

2
Antecedents

NUE is not only a very recent branch of urban economics, but it burst on the scene very suddenly in the late 1960s. Nevertheless its origins can be traced back in time to a much earlier date. Indeed, the closest historical antecedent of NUE is the urban version of von Thünen's agricultural land-use model developed in the 1820s. Although some early twentieth century urban-rent analysts (for example Hurd, Hoyt) have had some impact on urban economic theory in general and NUE in particular, the dominant influences have been the models of Alonso, Wingo, and Muth developed in the early 1960s. The contributions of these predecessors to NUE are analyzed in this chapter.

von Thünen

The standard model of urban land use that survives, in modified form, in NUE models derives from the early nineteenth century writings of J H von Thünen (1966). In fact, von Thünen's model refers to the spatial distribution of crops according to yield per unit area around a central town, and it remained for Isard (1956, page 200) to recognize that the model...

Table of contents

• Cover
• Title
• Copyright
• Preface
• Contents
• 1 What is 'New Urban Economics'?
• 2 Antecedents
• 3 The standard NUE model
• 4 Implications and extensions of the standard model
• 5 The monocentric city
• 6 The multicentric city
• 7 More complex residential location patterns
• 8 Locational interdependence
• 9 Towards dynamics
• 10 An optimum geography
• 11 Two residential location models
• 12 Alternatives to NUE
• 13 Political economy
• 14 Conclusion: are NUE models operational?
• References
• Name index
• Subject index