![Sample our Built Environment journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5926/TOC-Subject-Sample-2021-Built-Environment.jpg"})
Abstract
Problem: Reducing gasoline consumption could sharply curtail greenhouse gas emissions. Ongoing research seeks to document factors associated with green travel behavior, like walking and transit use.
Purpose: We seek to determine whether green beliefs and values are associated with green travel behavior. We measure whether residents of communities with environmentalist attributes drive less, consume less gasoline, and are more likely to commute by private vehicle. We explore several channels through which green beliefs and values may affect travel behavior and vice versa.
Methods: We drew our demographic, transportation, and built environment data from the 2000 Census of Population and Housing including the Public Use Microdata Sample and the 2001 National Household Travel Survey, and constructed our indicators of green ideology using voting records, political party membership, and data on hybrid auto ownership. We estimated ordinary least squares regression and linear probability models using both individual households and small areas as units of analysis.
Results and conclusions: We find green ideology is associated with green travel behavior. People with green values are more likely than others to be located in communities with high population densities and proximity to city centers and rail transit stations, which are attributes conducive to environmentally friendly travel. We also find that residents of green communities engage in more sustainable travel than residents of other communities, even controlling for demographics and the effects of the built environment. Green ideology may cause green travel behavior because greens derive utility from conservation or because greens locate in, or create, areas with characteristics that promote sustainable travel. We also discuss the possibility that green travel behavior may cause green beliefs.
Takeaway for practice: If greens self-select into dense, central, and transit-friendly areas, the demand for these characteristics may rise if green consciousness does. Alternatively, if these characteristics cause green consciousness, their promotion promises to increase green behavior. The implications of our finding that residents of green communities engage in more sustainable travel patterns than others depends on the causal mechanism at work. If greens conserve because they derive utility from it, then environmental education and persuasion may bring about more sustainable travel. Alternatively, if green travel behavior causes green beliefs, it is possible that attracting more travelers to alternate modes and reducing vehicle miles traveled may increase environmental consciousness, which may in turn promote other types of pro-environment behavior.
Research support: None.
Notes
Notes:
a. Each model includes a fixed effects term identifying the tract's MSA.
b. Based on the tract's Green Party share of registered voters and shares voting in favor of Propositions 12 and 13 in 2000.
Notes:
a. From the 2001 National Household Travel Survey.
b. Sample of households is drawn from those with heads aged 18 to 65 in metropolitan census tracts nationwide. The standard errors are clustered by metropolitan area. The omitted category is a white household whose zip code centroid is more than one mile from the closest rail transit station.
c. This is the representative's LCV score, on a scale of 0 to 100, in the 106th Congress. This variable has a mean of 36.5 and a standard deviation of 33.4.
Notes:
a. Each model includes a fixed effects term identifying the tract's MSA. Standard errors are clustered by congressional district. The omitted category is a census tract whose centroid is more than one mile from the closest rail transit station. The regressions are weighted by the count of commuters.
b. This is the representative's LCV score, on a scale of 0 to 100, in the 106th Congress. This variable has a mean of 36.4 and a standard deviation of 33.4.
Notes:
a. Each model includes a fixed effects term identifying the state in which the household was located. Standard errors are clustered by congressional district to control for the posibility of within-district correlation of the error terms.
b. Heads of households in the sample are those included in metropolitan PUMAs in the national PUMS data from the 2000 Census. The omitted category is a non-Black head of household.
c. This is the representative's LCV score in the 106th Congress on a score of 1 to 100, merged with the household data using the household's geographical identifier. This variable has a mean of 36.4 and a standard deviation of 33.4.
Notes:
a. Each model includes a fixed effects term identifying the MSA in which the household was located. The omitted category is a non-Black head of household. The standard errors are adjusted for within-PUMA correlation. There are 220 PUMAs in the sample.
b. The environmentalism factor is based on a PUMA's Green Party registration share, voting shares in favor of Propositions 12 and 13 in 2000, and the share of vehicels in the PUMA that are hybrids, based on 2007 registration data. It has a mean of zero and a standard deviation of one.
1. We recognize that there are suburbs and small towns that exhibit high levels of walking, biking, and bus ridership, despite the fact that they lack rail transit, have low densities, and are located great distances from major central business districts (CBDs). These include communities such as such as Davis, CA, Madison, WI, Ann Arbor, MI, and Bethesda, MD. Some, like Ann Arbor and Madison, are the principal cities in their own metropolitan statistical areas (MSAs), while others are included in larger MSAs (Davis in Sacramento, Bethesda in Washington, DC). In all, we do not claim our urban form variables perfectly capture all the elements that promote green travel behavior. But the evidence from the literature presented above indicates that the land use patterns and transportation architecture of suburbs and small towns are generally not conducive to green travel. The examples above are university towns, and thus attract intellectuals, who are perhaps disproportionately prone to environmental consciousness, and the young, who are often carless.
2. CBD definitions are taken from the Economic Census, Geographic Reference Manual (U.S. Census Bureau, 1982). They represent the agglomerations of census tracts that the local business leaders who were surveyed considered the centers of economic activity for each metropolitan region.
3. We acknowledge that our set of urban form variables is by no means exhaustive, and much of the existing literature examines built environment characteristics in greater detail; however, the broad geographic scope of our inquiry limits the urban form data available to us.
4. The NHTS collects VMT and derives vehicle fuel economy from vehicle information; it then calculates annual fuel consumption by dividing VMT by miles per gallon (Energy Information Administration, 2003).
5. For details on the construction of the dataset, see CitationBaum-Snow and Kahn (2005). Rail transit is defined as modern heavy or light rail rapid transit lines (such as San Francisco's BART), not vintage trolleys or commuter rail lines.
6. In the 1980s, when he was a U.S. Representative from Wyoming, Dick Cheney received a LCV score of 0% based on his voting on environmental bills over a two-year period.
7. Our factor analysis results indicated that the first factor explained over 90% of the variance. The individual environmental variables are highly correlated with each other. The correlation between voting in favor of Proposition 12 and 13 is .97. The lowest correlation (.25) is between the Green Party measure and voting in favor of Proposition 13. The variable's Cronbach's alpha has an acceptable value of .75. We acknowledge that there may be environmentalists who for independent reasons oppose all political referenda. In such communities, the Green Party membership would be high but the share voting in favor of green initiatives would be low.
8. “The National Green Pages is a directory listing nearly 3,000 businesses that have made firm commitments to sustainable, socially just principles, including the support of sweatshop-free labor, organic farms, fair trade, and cruelty-free products” (Green America, 2009, n.p.)
9. In results available on request, we have regressed each of our environmentalism indicators on our full set of urban form variables (distance to the CBD, population density and proximity to a rail transit station). The R 2 averages 0.5.
10. For each of these variables, we have top coded the outliers at the 99th percentile of the empirical distribution.
11. reports median regressions, robust to outliers.
12. Given the large geographical area covered by PUMAs, and given the fact that we do not know where households live within the PUMA, we do not create a measure of whether the household lives within one mile of a rail transit station. Using the PUMA's centroid to measure this distance would introduce measurement error.
13. All standard errors are clustered by PUMA.