Forty years of modelling rapid transit’s land value uplift in North America: moving beyond the tip of the iceberg

ABSTRACT

Identifying and measuring the land value uplift (LVU) impacts of rapid transit are important for a number of reasons. However, despite the general notion that rapid transit does confer positive LVU benefits, our comprehensive and critical review of more than 130 analyses across 60 studies completed in North America over the past 40 years finds significant heterogeneity in research outcomes, leaving many significant questions unanswered. Beyond high-level differences in study inputs, we argue that a fundamental source of variability is a lack of empirical specificity from the use of proximity as the dominant way in which LVU benefits are captured. This use of a proxy leads to the potential for omitted variables and unobserved relationships, and exposes previous work to the potential for misvalued results. To overcome this issue, we outline recommendations for future research, namely a recognition of relative accessibility and the possibility of LVU impacts from transit-oriented development. Incorporating measures related to these factors into LVU models can reveal their implicit prices, resulting in research that is more theoretically inclusive, empirically comprehensive, comparable, and able to provide important information to inform policy analysis and prescription.

KEYWORDS:

• Land value uplift
• rapid transit
• hedonic model
• land value capture
• accessibility
• self-selection
• transit-oriented development

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Christopher D. Higgins http://orcid.org/0000-0002-3551-7750

Notes

1 This sample consists of studies by Al-Mosaind, Dueker, and Strathman (1993), Alterkawi (1991), Armstrong (1994), Armstrong and Rodriguez (2006), Atkinson-Palombo (2010), Bajic (1983), Benjamin and Sirmans (1996), Bernick and Carroll (1991), Bernick, Cervero, and Menotti (1994), Bollinger, Ihlanfeldt, and Bowes (1998), Bowes and Ihlandfeldt (2001), Boyce, Allen, Mudge, Slater, and Isserman (1972), Cervero (1996; 2004), Cervero and Duncan (2002a; 2002b; 2002c), Cervero and Landis (1993), Chatman, Tulach, and Kim (2012), Chen, Rufolo, and Dueker (1998), Clower and Weinstein (2002), Damm, Lerman, Lerner-Lam, and Young (1980), Davis (1970), Dewees (1976), Diao (2015), Diao and Ferreira (2010), Dornbush (1975), Dubé, Des Rosiers, Thériault, and Dib (2011), Dubé, Thériault, and Des Rosiers (2013), Duncan (2008; 2011a; 2011b), Dyett et al. (1979), Falcke (1978), Farooq, Miller, and Haider (2010), Fejarang (1993), Ferguson, Goldberg, and Mark (1988), Garrett (2004), Gatzlaff and Smith (1993), Goetz, Ko, Hagar, Ton, and Matson (2010), Golub, Guhathakurta, and Sollapuram (2012), Grass (1992), Haider and Miller (2000), Hess and Almeida (2007), Hewitt and Hewitt (2012), Kahn (2007), Kim and Lahr (2013), Kittrell (2012), Knaap, Ding, and Hopkins (2001), Ko and Cao (2013), Landis and Loutzenheiser (1995), Landis, Guhathakurta, Huang, and Zhang (1995), Lee (1973), Lewis-Workman and Brod (1997), McDonald and Osuji (1995), McMillen and McDonald (2004), Nelson (1992; 1999), Nelson and McClesky (1990), Pan (2013), Perk and Catalá (2009), Perk, Bovino, Catalá, Reader, and Ulloa (2013), Petheram, Nelson, Miller, and Ewing (2013), Ryan (2005), Rybeck (1981), Seo, Golub, and Kuby (2014), Skaburskis (1982), Smith (1978), VNI Rainbow (1992), Voith (1991; 1993), Weinberger (2001), Weinstein and Clower (1999), and Yan, Delmelle, and Duncan (2012).

2 The earliest study known to the authors is that of Spengler’s (1930) comparison of assessed values in New York City.

3 Readers interested in a more detailed explanation of the model in general and applied to rapid transit specifically are directed to Brueckner (1987) and Debrezion et al. (2007).

4 Readers interested in a more detailed overview of the method and applications are directed to Malpezzi’s (2002) review.

5 Beyond the hedonic method, there are other methods with associated functional forms for estimating the value placed on environmental amenities, such as McFadden’s (1978) proposal to use of discrete choice methods grounded in random utility theory to model residential location. An integration of the bid-rent foundations of the AMM model and McFadden’s approach can be found in the work of Martinez (1992). However, to the best of our knowledge, no rapid transit LVU studies published to date have used these methods.

6 See Bowen et al. (2001) for a review.

7 Assuming that accessibility is effectively modelled, inter-modal comparisons may reveal interesting qualitative differences between modes.