Broken Windows‎ > ‎

Does urban sprawl cause obesity?

posted Jun 8, 2014, 2:15 PM by Joel Wood   [ updated Jan 27, 2015, 1:57 PM ]
If you listen to urban planners, urbanists, or public health officials in a major Canadian city these days, you will undoubtedly be told at some point that density, walkability, and urban design are of utter importance for public health. Intuitively this makes sense considering when we think of suburbs and exurbs, we think of highways, single family homes on large lots, big box stores (Walmart in particular), and shopping malls; the picture of car dependency. People will be less healthy because they will walk less because the suburban built environment heavily favours cars over pedestrians (as someone who grew up in the Vancouver suburb of Surrey, and then later on lived in Langley, I can attest to this). But is this intuition correct, and if not, what are the implications for public policy?

There are many studies out there looking at the relationship between sprawl and obesity. A study authored by researchers at St Michael's Hospital and the University of Toronto was published last January in PLOS ONE epitomizes much of the statistical analysis of the link between sprawl and obesity. The study analyzes cross-sectional data on Toronto neighbourhoods, and finds evidence that "Individuals in less walkable areas were up to one-third more likely to be obese or to have diabetes. Residential density and the availability of walkable destinations were each significantly associated with transportation and health outcomes." The study was widely covered by the Canadian media (here is a CBC story on the study). 

But what do the results of the study mean for public policy? Should walkable, dense communities be a public health priority? From this report from Toronto Public Health, they at the very least appear to believe it should be. However, the authors of the St. Michael's/UofT study are rightly careful to point out in the Discussion section that "the study was cross-sectional and therefore describes correlations but should not be used to draw causal inferences. For example, it is not possible to tell from these data if people walk more because they live in areas that are more walkable versus whether people who choose to walk move to these sorts of areas." What they are raising here is the neighbourhood selection effect. It may be that dense, walkable neighbourhoods are healthier because healthier people (or people predisposed to be healthy) choose to live in them over car-dependent neighbourhoods for unrelated reasons. Similarly, less healthier people (or even just people predisposed to being less healthy) may choose sprawling neighbourhoods. If the neighbourhood selection effect is driving the results, then a public policy encouraging density and walkability, especially at the municipal level, will not actually make anyone healthier; it will just encourage people who are less healthy (or predisposed to being less healthy) to move to other municipalities.

Luckily, some urban economists have studied the issue in a way that accounts for neighbourhood selection: they follow individuals over time. In an article published in the Journal of Urban Economics (ungated version here), Jean Eid (WLU), Henry Overman (LSE), Diego Puga (CEMFI), and Matthew Turner (University of Toronto) {Note: It has been my experience, that if there is an important urban policy question, Matthew Turner has probably done good work on it.} use a US data set that tracks 6,000 young adults over 6 years: address, weight, and other characteristics. The panel nature of their data, and the fact that most of those in the survey change addresses during the survey, allows them "to identify the effect of sprawl on weight, after controlling for individuals’ propensity to be obese, by looking at whether a given individual gains weight when they move to a more sprawling neighborhood or if they lose weight when they move to a less sprawling one". Their empirical analysis finds no effect of neighbourhood characteristics on weight after controlling for individuals' propensity to be obese. 

What does this mean for public policy? In my opinion it doesn't mean that we shouldn't be promoting dense, walkable neighbourhoods (there are other reasons these neighbourhoods are desirable); but it does mean we shouldn't be using public health to justify promoting density and chastising sprawl.