Household Choice and Urban Structure

A Re-Assessment of the Behavioural Foundations of Urban Models of Housing, Labor and Transportation Markets

Paul A. Waddell

• Publisher: Routledge
• Year: 2018
• Print ISBN: 9781138316324
• eBook ISBN: 9780429849589
• Edition: 1
• Topic: Physical Sciences
• Subtopic: Urban Planning & Landscaping

Published in 1997. The aim of this book is to explore urban modelling traditions, identify key limitations and contributions and to develop a more general model within a discrete choice framework. The scope of the effort is on household choices regarding residential location, workplace and housing tenure. It is the first systematic effort to analyze the structure and sequence of the choices made by households regarding residential location and workplace. The implications for urban theory, model development and policy analysis are substantial.

1 Introduction

The last several decades have witnessed a massive reorganization of urban American society, including a dramatic dispersal of population and jobs from central cities, the rise of 'edge cities,' profound demographic and social changes such as the prevalence of dual-earner households and a shrinking middle class. Urban policy has not kept pace with these changes, and neither has our understanding of the complex interactions between market forces and public policy in shaping the evolution of cities and regions. Both the research on urban development and the public policies intended to shape it have devolved into piecemeal and often incoherent frameworks.

Population and employment have suburbanized rapidly during the four decades since World War II, with widespread declines in the population of central cities (Berry and Dahmann, 1977). Suburbanization has accelerated as a result of massive public investments in freeways, rising affluence among large sections of the population, and white flight from the central cities. The urban landscape has been reshaped by new suburban office corridors along sprawling expressways, far from the central business districts which once dominated the urban form. Manufacturing employment has stagnated, while service employment has grown dramatically, with heavy consequences for the occupational demands and opportunities for different segments of the population, and strong spatial implications as the new service jobs cluster in high rise office buildings in the central business districts and throughout the suburban sprawl.

New ethnic concentrations have formed in metropolitan pockets, and the black ghettoes in declining central cities have become a seemingly permanent fixture. The ghetto underclass has faced diminishing access to, and contact with, the remainder of the urban economy. Travel patterns have lost their suburb to downtown orientation, and been replaced by dispersed suburb to suburb, crosstown, and reverse-commuting patterns which frequently congest the beltways in the suburbs as much as the radial freeways emanating from the city center. Suburbs have mushroomed and multiplied until they fragment the urban landscape and preclude further expansion of the central cities, limiting the capacity of central cities to compete for decentralizing jobs and households.

The urban labor market, the urban housing market, transportation policy and infrastructure, and individual and collective attitudes such as racial discrimination all interact to form a physical, economic, and social map of the metropolis. Perhaps one reason so many urban policies have failed in the past relates to their failure to address this systematic set of spatial interactions in an informed fashion, leading to fragmentary and incomplete, or worse contradictory policies for urban development.

Similarly, the academic study of the urban environment has often failed to adequately integrate the relevant disciplines to focus on urban problems in a systematic manner. The models of urban dynamics which have been developed in the fields of urban economics and sociology have added significant insights into the nature of urban problems, but each provides only a partial perspective, and few adequately incorporate the spatial context in which urban processes transpire. Nor do current modeling efforts adequately describe public policy dimensions. Considering the degree to which the urban landscape is shaped by public investments, this is a serious shortcoming.

Several efforts have been made recently to cross disciplinary boundaries to attempt to develop quantitative urban models of the interactions between the labor market, housing market, transportation system, and public policy in a spatial context. While these tend to be well grounded in urban economic theory, there is a shift towards the implementation of such models using random utility theory, which is a stochastic extension of consumer economic theory. These models are typically operationalized with multinomial or nested logit formulations.

Among the more significant recent models of this type are the housing and tenure submodels of the Harvard Urban Development Simulation (HUDS) model described by Kain and Apgar (1985), and the Chicago Area Transportation and Land Analysis System (CATLAS) developed by Anas (1982). The HUDS model uses a multinomial logit model of the demand for bundles of housing attributes and neighborhood location for 96 different household types by racial, socioeconomic, and demographic category, given workplace location and travel costs. CATLAS develops a simultaneous equilibrium nested logit model of travel mode to work and residential location choice, given workplace and travel cost.

The model developed in this study is an attempt to extend this line of research by focusing on several related choices made by individual workers. The choice of a job given the location of firms, the choice of residence location given the spatial distribution of housing supply, the choice of housing tenure, and the choice of travel mode to work are all related decisions made by employed householders. These decisions can easily be seen to be related, but the exact nature and sequence of the decision making process is less clear.

The major urban models which were developed in the past several decades, including those developed from the monocentric model of Alonso (1964) and the Lowry (1964) gravity model, have assumed workplace to be exogenous in determining residential location. Hamilton (1982), however, has found that actual commuting distance is about eight times larger than would be expected from the monocentric model, indicating a large degree of 'wasteful commuting'. One of the main reasons for this result, according to Hamilton, is that residential relocation costs are significant, and may in many cases lead to workplace location decisions being made from predetermined residential locations.

Similarly, Lirineman and Graves (1983) have argued that 'job search and residence decisions are intimately intertwined over both short and long distances', and Weinberg (1979) that individuals can adjust accessibility by adjusting workplace location or residence location or both. Simpson (1980,1986) has extended these arguments by estimating a simultaneous equation model of residence and workplace location using microdata for Toronto. Although individuals would not be expected to make simultaneous decisions regarding their residence and workplace locations (Gordon and Vickerman, 1982), some individuals will make workplace decisions based of predetermined residence location while others make residence decisions on the basis of predetermined workplace locations.

The degree to which residence location is driven by workplace location, or the converse, may also vary by household relationship, tenure, ethnicity, and socioeconomic status. Beesley and Dalvi (1974) have argued that residential choice is primarily the decision of the household head, with the likely implication that secondary workers in the household choose their workplace on the basis of a predetermined residential location. Homeowners face higher relocation costs than renters, and would be expected to be more likely than renters to choose their workplace on the basis of their residential location. To the extent that blacks face discrimination in housing choice, their residential location opportunities are restricted, and they are more likely to be forced to choose a job based on their residential location. Schwartz (1973) has found that more educated workers adopt a larger job search radius, confirming that higher socioeconomic status confers greater flexibility in the choice of both residential and workplace location.

This study uses multinomial and nested logit modeling as tools to examine this interdependent decisionmaking process in detail, for white, black, and Hispanic workers. The resulting model of the choice of workplace location, residence location, housing tenure, and travel mode provides a more systematic view of the interrelationships between, and spatial aspects of, labor supply, housing demand, and transportation service demand. The public policy areas impacted by these areas include transportation and land use planning, housing, labor, economic development, and racial discrimination.

The model is estimated using a special tabulation prepared by the U.S. Census Bureau Division of Journey to Work and Migration specifically for this study, using 1980 data for the Dallas-Fort Worth SMS A. The data contains detailed commuting patterns and housing choices by travel mode, and by detailed racial, demographic, and socioeconomic worker categories.

2 Theoretical background: models of urban spatial structure

Several strains of theoretical development are relevant to the objectives of this research, and will be discussed at some length in this chapter. Although this background only skims the surface of the vast literature which exists in these areas, it will suffice to provide a foundation to pursue a strand of research which integrates several of these branches of research. The principal components to be discussed in this chapter are 1) the development and extension of urban simulation models, particularly gravity models, 2) the monocentric model of urban economic theory, 3) random utility theory and discrete choice modeling, and 4) social and urban ecology.

The development of theoretical and quantitative models of the spatial characteristics of the urban housing and labor markets, and the resulting pattern of intraurban commuting, has evolved along many diverse channels. The land-use forecasting models developed as inputs to the transportation planning process are perhaps the most widely implemented models of urban spatial activity. Most of these models are grouped into a class of models known as gravity models, and have been employed more for their computational tractability than their underlying theoretical foundations. These models have been stretched to include more elements of household behavior, but are still often criticized for their lack of a solid foundation in behavioral theory.

Other urban simulations have been developed to analyze a variety of issues related to urban spatial structure, including several housing market models with a high degree of household and housing detail and some spatial detail. Typically, these models are better rooted in urban economic theory than the gravity models, but tend to be designed for special purpose applications and are extremely complex.

The field of urban economics has developed over the ...

Table of contents

• Cover
• Half Title
• Title
• Copyright
• Contents
• List of figures
• List of tables
• Acknowledgments
• 1 Introduction
• 2 Theoretical background: models of urban spatial structure
• 3 The policy interface
• 4 The region, the data, and the model
• 5 Results and interpretation
• 6 Conclusions and directions for future research
• Bibliography