eres2010_130_Henneberry_MODELLING_THE_RELATI(12)

The paper describes the development and application of physical-financialmodelling techniques to the analysis of relations between development design –covering the broad characteristics of a scheme, such as land use mix, developmentdensity and built form – and financial viability. It is divided into two parts

argument (Fraser, 1993; Evans, 2004; Wyatt, 2007; Ball et al, 2008). For a site of a fixed size, development density is represented by the amount of floor space built on it. At low densities, the marginal cost (MC) of an extra unit of floor space falls. Consider building two stories instead of one: the cost of the foundations varies little, but the roof covers (almost) twice as much (usable) floor space. However, after a certain point, marginal costs begin and continue to rise because of the need for stronger foundations, more services such as lifts and so on.

Conversely, marginal revenue (MR) declines as development density increases. Occupiers derive less utility from property on the site as local environmental quality dwindles – because of less open space, more noise and pollution and so on - and the accessibility of the upper floors in higher buildings decreases. The optimum development density is X, where MR equals MC, giving a site value of PQY. However, by "creating value through the appropriate densities (sic), public space, uses and distribution of buildings" (Roger Evans Associates, 2007, page 110) good urban design can shift the marginal revenue curve upwards (to MR1), allowing higher density development (to X1) that is financially viable and increasing the site value to PQ1Y1.

Every element of this argument is open to question. Little is known of the positions or slopes of the MC and MR curves, or of the point of inflection of the former. Flanagan and Norman (1978) suggest that the marginal cost starts to increase at six stories, whereas Picken and Ilozor (2003) suggest that the turning point is at 35 stories. Chau et al (2007) found that for flats with views, marginal revenue increased exponentially with an increase in floor level and that the optimal building height was higher for a site with a ‘better’ external environment. Eppli and Tu (1999) found that house prices were higher in higher density, New Urbanist schemes than in conventional suburban developments; although this was in a very low-density environment. In contrast, Song and Knaap (2003) and Waasmer and Bass (2006) show that residents pay less for houses in denser, more central neighbourhoods.

What such studies indicate is “… that the curves …[in Figure 4]… merely represent general trends in marginal revenue and cost. In any individual development, the shapes of the curves would vary …” (Fraser, 1993, page 233). These circumstances provide a strong justification for undertaking more research on the relation between development design and viability at two levels. The first is the site level. The modelling system may be used to capture the particularities and varieties of physical form and financial structure. Development density "is a product of design, not a determinant of it" (Llewelyn-Davies, 2000, page 46). Building type and height, block size, the positions of buildings relative to one another and the distribution and quality of open space will all affect perceptions of density. Such perceptions, in turn, affect the demand for and value of development. The second is the level of analytical generalisation. Modelling techniques allow the wider implications of the replication of various sets of site-specific relations to be explored from the bottom up. Methodological choice is not limited to the breadth-depth trade-off between aggregate quantitative and qualitative case study approaches. Different types of development project display different links between design and value. There is not one but a series of optimum development densities, the determination of which depend on sites’ characteristics and contexts.