On the Issue of Orderedness in Social Accounting Models of Residential Change

Abstract

Residential Change in a Social Accounting FrameworkParticular attention has been devoted to changes occurring within the existing housing stock of a metropolitan area, since the standing stock provides the physical space for urban activities. From the urban planner's point of view, the important questions about neighborhood change deal directly with compositional changes in occupancy of the housing stock. In this manner, the planner is more concerned with identifying the characteristics of households occupying a neighborhood at a given time period rather than monitoring the movements of individual households whose destinations may lie outside the neighborhood. Therefore, by focusing on a set of dwelling units with their attendant household composition, modeling endeavors accounting for the shifts in the characteristics of a sequence of household types by dwelling unit would be invaluable to planners as a means of discerning household compositional shifts in the housing stock (Moore and Gale 1979).Social accounting models of residential changes have been endorsed recently in lieu of conventional microeconomic attempts at uncovering the exchange mechanisms between households and dwelling units (Byler and Gale 1978; Moore and Gale 1979;Stone 1975) The latter's difficulty with (a) identifying extremely localized pockets of population movement, (b) providing current and reliable predictions on household/dwelling unit composition in housing submarkets, and (c) formulating the housing process solely within a market-based model reflect the foremost factors for this shift in modeling attempts of the housing process. In its primitive form, accounting models of the housing process represent an extension of the "models of data" concept proposed by Suppes (1962) and elaborated by Moore and Gale (1974). These models are useful for monitoring the sequence of states through which elements of the system pass. For example, such models may be used to "account" for the spatial pattern of change in (a) housing stock composition caused by construction, demolition, or other types of structural modification; (b) household composition due to migration, formulation, or other types of consumer household change; and (c) the matching process between the two elements.General modeling focus on these activities has been structured within a contingency-table formulation in which the table represents longitudinal time sequences of household types occupying specific classes of dwelling units. These tables may be multidimensional in form where both elements, household and dwelling unit type, are represented by a set of attributes depicting various properties of the two elements. Although this contingency-table framework has been used for housing inventory and descriptive purposes, more parsimonious "accounting" models have been proposed.