The “conventional wisdom” holds that neighbourhood effects are crucially important in shaping the development and life prospects of children. In particular, it is widely believed that growing up in very poor neighbourhoods has powerful negative effects on the cognitive and social development of children.
Whether or not “neighbourhood effects” influence the development of young people is a question that has important implications for public policy. It is a widely held opinion that growing up in neighbourhoods where poverty is common has strong negative effects on the social and cognitive development of children. This has led many policymakers and activists to argue in favour of government interventions designed to target entire communities rather than spending resources directly on individuals. Through projects of community redevelopment, some argue that government can help eliminate the sources of negative neighbourhood effects, thereby removing a major impediment to future social and economic success for the children who grow up in poor communities. The wisdom of this “community-based” approach to facilitating the development of children largely hinges on whether or not negative neighbourhood effects from living in poor communities actually exist.
Indeed, until recently, the overwhelming majority of the research surrounding this question found that there are important, negative effects associated with living in poor, crime ridden neighbourhoods. In a recent review of the relevant literature on the subject, however, Professor Phil Oreopoulos argues that the actual evidence surrounding neighbourhood effects is actually quite ambiguous.
Professor Oreopoulos’ review of the existing literature surrounding neighbourhood effects begins with a discussion of ethnographic studies, in which researchers have frequently argued in favour of the existence of negative neighbourhood effects. Oreopoulos points out that this approach is imperfect for understanding the supposed phenomena of neighbourhood effects generally because it necessarily focuses on very specific examples that are of particular interest to the researcher. Oreopoulos also notes that this approach, which relies on case studies and interviews, always has a strong potential for bias, since researchers may consciously or unconsciously be influenced by their preexisting beliefs when deciding what material to include and exclude from their reports. Although Oreopoulos states that these ethnographic studies provide compelling examples of the ways that neighbourhoods can effect well being, they cannot accurately determine the causal relationship between neighbourhood effects and welfare outcomes.
A second approach that has been used to study and attempt to measure neighbourhood effects is the use of multivariate regression analysis. These studies have also, generally, found significant negative effects associated with living in a poor neighbourhood. In his literature review, however, Oreopoulos argues that claims for the existence of neighbourhood effects on the basis of this methodological approach are unconvincing. Oreopoulos writes that, due to the enormously diverse motivations that impact people’s choices about where to live, it is impossible to eliminate the possibility of omitted variable bias through the identification of control variables. Due to the likelihood that unidentified characteristics, observable or non-observable, exist which impact both outcomes and neighbourhood choice, Oreopoulos argues that estimates of neighbourhood effects based on multivariate regression should be met with skepticism
While ethnographic studies and multivariate regression are unable to clearly establish the causal relationship between neighbourhoods and child development outcomes, Professor Oreopoulos writes that experiments with random or quasi-random assignment hold the potential to dramatically improve our understanding of neighbourhood effects. Oreopoulos describes three quasi-experimental studies which have examined neighbourhood effects in recent years. Of these, Professor Oreopoulos’ own study from 2003, “The Long-Run Consequences of Living in a Poor Neighbourhood,” was the only Canadian example. In this study, Oreopoulos took advantage of a “natural experiment” created by a Toronto policy that was in place prior to the early 1980s. This policy awarded subsidized housing units to public housing applicants as the units became available. These units were located in different locations across the city. Some applicants were placed in large public housing projects where poverty rates were extremely high. Others were assigned to much smaller complexes of townhouses in other parts of the city nearby to middle-income neighbourhoods. Because of this variation, city-block and census tract characteristics, which are the units usually employed to study neighbourhood effects, varied across the projects.
Since public housing recipients were assigned to one neighbourhood or another randomly, depending on what sort of unit became available when they reached the top of the list, this policy had the effect of creating a natural experiment that could be used to study neighbourhood effects. In this natural experiment, the families assigned to very poor neighbourhoods could be directly compared to the individuals placed in wealthier neighbourhoods.
Due to the random assignment, we can safely assume that there existed no other difference, on average in the characteristics of the groups being compared other than their neighbourhood assignment. If the children in the group assigned to wealthier neighbourhoods achieved significantly better developmental outcomes on average than those in the poor neighbourhoods, the difference between the two groups could be safely attributed to the neighbourhood effect.
In order to study the development outcomes of the children in this natural experiment, Oreopoulos used administrative data that tracked the children who grew up in these projects over the course of many years, until they were more than thirty years of age. Oreopoulos was able to compare long-term measures of total income, wages, unemployment and public assistance of otherwise similar groups of children who were raised in different neighbourhoods. Perhaps surprisingly, Oreopoulos’ analysis of the data discovered that there was no difference in these outcomes across the different types of projects.
The most important limitation of studies such as that performed by Professor Oreopoulos is that they can only inform us about the neighbourhood effect, or lack thereof, in a particular time and place. As Oreopoulos notes, there may be significant differences between poor communities in Toronto and poor communities in the United States, or even elsewhere in Canada. It is possible that while growing up in Regent Park in Toronto does not carry a negative neighbourhood effect, growing up in a different poor community does.
For example, if another community, even one with an identical average income to Regent Park, is heavily infested with crime and gangs, it is possible that living there would have a negative neighbourhood effect, even if Regent Park does not. Certainly, Oreopoulos’ study provides very strong evidence in support of the argument that neighbourhood effects are much less important than is widely believed. However, in order for it to be conclusively demonstrated that neighbourhood effects simply do not matter for child development, more studies would need to be performed in a wide variety of contexts.