International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Endogenous treatment of residential location choices in transport and land use models: introducing the MetroScan framework
Chinh Ho, David Hensher

Last modified: 28 March 2017


A recognition that transport and land use mutually influence each other has produced a growing interest in integrating transport and land use models for metropolitan planning process. Since Lerman’s (1976) pioneering work to link the long-term decision of residential location with shorter-term decisions of workplace, vehicle ownership, and commuting mode in a multinominal logit model, many attempts have been made to extend this modelling approach to examine the interactions between transport and land use. Examples range from Abraham and Hunt’s (1997) model that captures the interactions between location choices (home and work) and mode choices through a nested logit model with variable nesting structure, to the UrbanSim framework that uses travel demand as an input into land use models that simulate the development of housing and labour markets (Waddell, 2000), to recent advancements of activity-based models in which the interdependencies between long-term and shorter-term decisions are captured through a series of discrete choice models that are linked together via the accessibility measure (Bowman and Ben-Akiva, 2001; Bradley and Bowman, 2006; Davidson et al., 2011).


Discrete choice models have created additional opportunities for modelling the interaction between transport and land use. However, most operational activity-travel demand models have not yet taken advantage of these opportunities for modelling residential location choice. For example, the two major operational frameworks, namely the DaySim (Bradley and Bowman, 2006) and the CT-RAMP (Davidson et al., 2010) family of activity-based modelling, treat residential location as exogenous to other travel-related decisions. Specifically, the choice of residential location is synthesised as opposed to being modelled endogenously in the modelling system. Synthesising where people live may be sufficient for understanding how residential location influences short-term travel behaviour, but this offers no clues as to how day-to-day travel experiences and job mobility may factor into longer-term household decisions to change residential location.


Within a random utility maximisation modelling framework, the current study develops a residential location choice model as part of an integrated transport and land use modelling system, called MetroScan – a quick scanning tool to evaluate transport and land use initiatives, including benefit-cost analysis and economic impact analysis. Figure 1 shows an overall structure of MetroScan demand and supply models. MetroScan includes not only passenger modules, but also firm location, light commercial vehicle (LCV) and freight modules. The firm location models provide input into the workplace location choice while the passenger module interacts with the LCV and freight modules on the network as they all compete for network capacity.


We focus herein on the residential location choice, which is a major household decision that determines not only the accessibility of each household member to their daily activities but also the household’s budget for other consumptions such as vehicle and dwelling ownership. The residential location choice model is structured above the workplace location choice in light of empirical evidence which suggests that 80% of households choose residential location first and then household workers choose their workplaces conditioned on residential location (Waddell et al., 2007). In addition, the household choice of residential location conditions other choices of tenure and dwelling type, household fleet size, places for non-work activities, and work practices (e.g., telecommute vs. compressed work week). Short-term decisions of travel mode and time of day travelled also influences residential location choice but only indirectly via the medium-term decisions of vehicle ownership, work and non-work location choices.


Figure 1. Overall framework of MetroScan demand and supply models


Modelling residential location choice within an integrated model of transport and land use is potentially the best way to investigate the interactions of residential location with other key decisions such as workplace, tenure and dwelling type, vehicle ownership and daily travel patterns. However, this approach requires that models describing these key decisions be estimated first to obtain the logsums for use in modelling residential location choice. All modules of MetroScan have been estimated and some detailed models can be found in previous papers (Ho and Hensher, 2014; 2016). This paper sets out the residential location choice model for Sydney residents and the empirical evidence, complete with all of the interconnected elements. To this end, a full implementation of MetroScan on a sample of synthetic households is conducted to study the impact of a recently announced Sydney Metro on residential relocation. These households are synthesised in such a way that they are representative of the population in terms of household size, household structure, number of household workers, occupation and work industry, age, income and other demographics (Ellison and Hensher, 2016).




Abraham, J. & Hunt, J. 1997. Specification and Estimation of Nested Logit Model of Home, Workplaces, and Commuter Mode Choices by Multiple-Worker Households. Transportation Research Record: Journal of the Transportation Research Board, 1606, 17-24.

Bowman, J. L. & Ben-Akiva, M. E. 2001. Activity-based disaggregate travel demand model system with activity schedules. Transportation Research Part A: Policy and Practice, 35, 1-28.

Bradley, M. & Bowman, J. 2006. SACSIM/05 Activity-Based Travel Forecasting Model for SACOG: Featuring DAYSIM – the Person Day Activity and Travel Simulator. Technical Memo Number 4: Mode Choice Models. Prepared for Sacramento Area Council of Governments (SACOG).

Davidson, B., Vovsha, P. & Freedman, J. New Advancements in Activity-Based Models.  Australasian Transport Research Forum 2011, 28 - 30 September 2011 Adelaide, Australia.

Davidson, W., Vovsha, P., Freedman, J. & Donnelly, R. CT-RAMP Family of Activity-Based Models.  Proceedings of the 33rd Australasian Transport Research Forum (ATRF), 29 September – 01 October 2010 Canberra, Australia.

Ellison, R. & Hensher, D. 2016. The TRESIS approach to spatial microsimulation. In: LOVELACE, R. & DUMONT, M. (eds.) Spatial Microsimulation in R. London: Chapman and Hall.

Ho, C. Q. & Hensher, D. A. 2014. Housing prices and price endogeneity in tenure and dwelling type choice models. Case Studies on Transport Policy, 107-115.

Ho, C. Q. & Hensher, D. A. 2016. A workplace choice model accounting for spatial competition and agglomeration effects. Journal of Transport Geography, 51, 193-203.

Lerman, S. R. 1976. Location, housing, automobile ownership, and mode to work: a joint choice model. Transportation Research Record, 610, 6-11.

Waddell, P. 2000. A Behavioral Simulation Model for Metropolitan Policy Analysis and Planning: Residential Location and Housing Market Components of Urbansim. Environment and Planning B: Planning and Design, 27, 247-263.

Waddell, P., Bhat, C., Eluru, N., Wang, L. & Pendyala, R. 2007. Modeling interdependence in household residence and workplace choices. Transportation Research Record: Journal of the Transportation Research Board.


Conference registration is required in order to view papers.