1.
Introduction
Increasing mobility causes the authorities a lot of worry. Commuter traffic is the major
culprit here. In line with a number of other European countries the Dutch government wishes
to limit commuter traffic through spatial policy (Vinex 1991; Van der Knaap & Van der Laan
1993). Also in the US, similar policies have been proposed (see for a discussion: Cervero &
Wu, 1997). On the one hand, policies should be attuned to the supply of labour, the places of
residence, and on the other hand, the focus should be on location policies in relation to labour
demand. Both aim at the simultaneous concentration of the place of work and residence.
Bringing jobs and housing together may lead to a jobs-housing balance and by this to shorter
commuter distances and less traffic.
This issue has a theoretical basis. What matters here is the appropriateness of models
used in urban analysis. Traditionally the analysis of commuting used the monocentric urban
model (Alonso 1964; Hamilton 1982; Yinger 1992). Particularly in urban economics this
model, even in its simplist form, is still surprisingly dominant as a research paradigm (Bourne,
1995; Gordon & Richardson, 1996a). The monocentric model starts from the individual or
the household, assuming that only one of the household members has a job. Households select,
within a daily urban system, a place to live, given the location of employment in the city
centre. Spending preferences of households are expressed in the indifference curve: the benefit
a household derives from different combinations of goods. Households weigh the pros and
cons of certain combinations of costs of housing (land price), commuting and other goods.
The same utility function applies to all households. The maximum amount a household can
spend is determined by the income. In the model this income is given and spent on housing,
commuting and other goods and services. The expenses for living concern the cost of a
certain location, in which the price of land is the major variable. The longer the distance to
the city centre, the higher the cost of commuting. Together with the utility function, the
budget equation determines the place where a household wishes to live.
This basic monocentric model is challenged from at least two sides (see Richardson
1988; Waddell 1993; Boarnet 1994). The first criticism is the emergence of polynodal
employment locations. The monocentric urban model is based on the city as a daily urban
system (DUS) with an urban core and a suburban area surrounding it. The DUS became a
synonym for a local urban labour market. The boundaries of this labour market are based on
the hierarchical-nodal principle: a dispersed labour supply directed at a central location of
labour-demand. Increasingly this conceptual framework is no longer valid. The basic model
loses its explanatory power because industry deconcentrates resulting in the emergence of
subcentres and by this of polynodal urban regions (see Erickson 1983; Goodschild & Munton
1985; Law 1988; Kumar 1990; Berry & Kim 1993; Clark & Kuijpers-Linde, 1994; Gordon &
Richardson, 1996a). There are even studies which point at a next stage in the
deconcentration process: the dispersed metropolis (Gordon & Richardson, 1996b). The
deconcentration is caused by economic and social-cultural reasons and, particularly in Europe,
by governmental interference. Because the traditional hierarchic structure changed into a
horizontal one, the urban system cannot be described anymore with the help of the hierarchic
model. The traditional duality between centre and suburb disappears and a multi-centred urban
area develops. In consequence the traditional nodal model becomes less attractive. This is the
reason why recent urban research and models start from the proposition of polynodality. But
also in these new models the propositions of the basic model in relation to land rent and costs
of commuting are kept. Also maintained is quite always the proposition that the ‘residential
choice occurs in an exogeneously system of workplace location’ (Berry & Kim 1993, p.3)