Cover Image

Not for Sale

View/Hide Left Panel
Click for next page ( 76

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 75
64 I N N O VAT I O N S I N T R AV E L D E M A N D M O D E L I N G , V O L U M E 2 optimal allocation of land in the sense that each house- Residential location choice was modeled via a multi- hold chooses a home that most satisfies it while develop- nomial logit framework. The random utility was speci- ers and land owners maximize profits and rents. The fied as follows: spatial distribution of households and the equilibrium 2 home prices are endogenously determined as the out- UNITPi SQFOOTi UNITPi SQFOO OTi U h i = 1 + 2 + 3 come of a housing-market mechanism involving land HHINCh HHINC h and transport. Nh SQFOOTi + 4 1, 000 + 5 ( VOTT h TThin ) n =1 DATA AND METHODS Nh This section describes the data used to calibrate the loca- + 6 ( DISTWORK ) + DISTMALL hin 7 i + hi n =1 tion choice model and to reach single-family housing market equilibriums. Both procedures were coded in where GAUSS matrix programming language (Aptech Systems 2003). Uhi = random utility of household h for choosing home i, 1, 1, . . . , 7 = parameters to be estimated, Location Choice Model Nh = number of workers in household h, VOTTh = household's approximate value of Bina and Kockelman (2006) undertook a survey of travel time, Austin movers in 2005. Sampling half of Travis County's TThin = network commute time for worker recent [with "recent" meaning within the past 12 months n in household h when residing in (before the sampling date and start of the survey)] home home i, buyers, responses were obtained from over 900 house- DISTWORKhin = corresponding Euclidean distance, holds, or roughly 12% of those buyers. The data set con- and tains comprehensive information on household dem- hi = random component assumed to be ographics, housing characteristics, reasons for reloca- independent identically distributed tion, and preferences when facing different housing and (IID) Gumbel, across households h location choice scenarios. and their alternatives i. Commute distance and cost have a bearing on one's residential location choice (e.g., Van Ommeren et al. For model calibration, each household's choice set con- 1999, Rouwendal and Meijer 2001, Clark et al. 2003, sisted of 20 alternatives: 19 randomly drawn from the Tillema et al. 2006). The GIS-encoded addresses of homes pool of all homes purchased by respondents in the and workplaces, accompanied by roadway network data, recent mover survey plus the chosen option. These provide a direct measure of commute time (with com- model results are shown in Table 1. The model indicates mute time calculated by using Caliper's TransCAD soft- a concave relationship between strength of preference ware for shortest travel-time path under free-flow (systematic utility) and the ratio of home price to annual conditions) for all potential locations. Because value of income. The parameter values on the ratio and its travel time (VOTT) was not directly available in the data squared term suggest that more expensive homes are set, it was approximated as the average wage (with part- preferred when the ratio is less than 1.7, becoming less time employed persons assumed to work 1,000 h/year, attractive as the ratio of price to income exceeds this while full-time employed persons were assumed to work threshold. 2,000 h/year) of each household's employed members Larger homes, of course, are more desired, with and was assumed to be equal over the household's SQFOOT increasing the likelihood of a home's selection, employed members. In addition to work access, each everything else constant. The negative signs associated potential home's (Euclidean) distance to the nearest of the with commute costs and Euclidean distances to workers' region's largest 18 shopping centers (DISTMALL) helps workplaces support the notion that households favor explain the impact of shopping access on location choice. homes closer to their employed workers' jobs. Major Furthermore, household annual (pretax) income mall access, however, was not favored; perhaps the (HHINC), home size (SQFOOT), and housing prices per potentially high volumes of traffic and congestion in the interior (built) square foot (UNITP) help explain the bal- vicinity of major shopping centers offset any possible ance of home affordability [where price (equal to UNITP access gains. Other forms of shopping access may be SQFOOT) is divided by annual income] and house- desired but require geocoding of far more smaller holds' preferences for larger home sizes. shopping.