Commute mode diversity and income inequality: an inter-urban analysis of 148 midsize US cities

ABSTRACT

Little is known about the relationship between multi-modal transportation environments and aspects of sustainable urban development, such as reduced income inequality and affordable housing. This study, adapted from Molotch and Appelbaum's inter-city differential method, studied 148 semi-isolated mid-size cities. Using U.S. Census data from 2013, we found that increased diversity in commute modality is associated with less income inequality between white and African-American households, as well as between men and women. Commute mode diversity is also associated with higher earnings for white women and African-American men. Our study also shows that cities with more commuter mode diversity are associated with higher home property values and affordable rental markets compared to automobile dependent cities. These results undercut the notion that increasing automobile ownership is a reasonable policy response to urban poverty, and suggest that sustainable transportation policy can produce positive economic gains for cities. This work adds to the growing literature identifying fundamental differences between multimodal and automobile dependent cities.

KEYWORDS:

• Automobile dependence
• multimodality
• income inequality
• urban planning
• sustainable urban development

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Chad Frederick http://orcid.org/0000-0001-8861-0708

Notes

1. Researchers are rightfully concerned about the ecological fallacy, but seem less concerned with its opposite: the atomistic fallacy. This occurs when the characteristics of cities and neighbourhoods are assumed to result from the characteristics of smaller units, such as Census blocks or individuals, respectively (see Schwartz 1994, Curtis and Rees 1998, Diez Roux 2001).

2. CMD minus carpooling and working from home: All wage-based relationships hold when the commute mode variable excludes working from home (WFH) and carpooling (changes noted below). One exception is the model for the difference between white and black median household income. In that case, both density and CMD stop contributing to the model; this shows car-pooling is important factor in black median household incomes.

•    (1) Difference between Male and Female Earnings. No remarkable changes: median household income becomes less explanatory (β = .401, p < .001) as does CMD (β = −.320, p < .001).

•    (2) Difference between White Male and Female Earnings. Again, median household income loses some explanatory power (β = .325, p < .001), as does CMD (β = −.324, p < .001).

•    (3) Difference between Black Male and Female Earnings. College education becomes marginally significant (β = −.176, p = .094). Latitude becomes more explanatory (β = −.258, p = .004), as does median household income (β = .342, p < .001), and CMD (β = .239, p = .015).

•    (4) Difference between White and Black Male Earnings. CMD changes slightly (β = −.181, p = .028).

•    (5) Difference between White and Black Female Earnings. Latitude becomes less important (β = −.217, p = .013), while CMD becomes slightly more important (β = .273, p = .001).