توزیع درآمد همسایگان: تبعیض نژادی اقتصادی و اختلاط در محله های شهری ایالات متحده

The paper describes within-neighborhood economic segregation in US metropolitan areas in 1985 and 1993. It uses the neighborhood clusters of the American housing survey, standardized by metropolitan area income and household size, to explore income distribution within neighborhoods at a scale much smaller than the census tract (a representative sample of households or ‘kernels’ and their 10 closest neighbors). Joint and conditional distributions portray neighbors’ characteristics conditional on the kernel’s housing tenure, race, and income. The paper documents both significant income mixing in the majority of US urban micro neighborhoods and the extent of income mixing within neighborhoods of concentrated poverty.

The distribution of income in residential neighborhoods matters. Since Alfred Marshall, economists have known of the role of nonmarket interactions and externalities in cities, and we know that for firms urban diversity increases the value of those interactions (Quigley, 1998). The same is true for households: a rapidly growing literature in economics documents the importance for households of nonmarket social interactions and externalities in cities (Glaeser, 2000). Income homogeneity or diversity is one of many dimensions of neighborhood social interactions yet unlike racial segregation, economic segregation as a feature of US neighborhoods attracted little attention from economists until recently. Nonmarket social interactions occur whenever one household’s characteristics affect its neighbors’ behaviors or socioeconomic outcomes. For example, if neighbors provide role models (positive or negative) or labor market connections then the productivity of investment in children’s education may be affected by a neighborhood’s income distribution, (Durlauf, 2003). The value of neighborhood interactions has attracted policymakers’ attention and led to policy initiatives intended to take advantage of positive externalities associated with mixing households of different income levels in neighborhoods.1 Yet we know surprisingly little about the degree of economic mixing or segregation within US neighborhoods, certainly much less than we know about racial segregation. Using a representative sample of US urban households and their immediate neighbors, the American housing survey’s neighborhood clusters data, this paper provides a portrait of the distribution of income and other socioeconomic characteristics among the immediate neighbors of a random sample of US households in 1985 and 1993. There is no unique definition of a neighborhood and economic segregation in neighborhoods can be viewed at many scales (Ellen, 1999 and White, 1987). What we know about income distribution within US urban neighborhoods has been limited by the data available. The most disaggregated data that US studies have used are decennial census data for census tracts (with an average population of 4000): mean and median family and household income, per capita income and poverty rates. In household-level micro data sets spatial detail is concealed to preserve respondents’ privacy. The smallest geographical identification is metropolitan area for the American housing survey and PUMA’s (with a population of 40,000 or more) for the Census Public Use Microdata. Because spatial detail was not available in these household level data sets, it was impossible to use them to analyze income distribution for smaller areas. Yet many neighborhood interactions take place at the scale of neighbors on the same block or in the same apartment building, rather than in the neighborhoods of several thousand people represented by census tracts.2 This paper presents the results of an empirical study of income mixing in neighborhoods of US cities using the neighborhood clusters data, a relatively neglected feature of the American housing survey. We briefly discuss theoretical and policy issues immediately below. Section 2 provides a detailed description of the neighborhood clusters data of the American housing survey (AHS) and of the measures of income used here, notably the HUD-adjusted median family income (HAMFI). Section 3 discusses alternative measures of income mixing and summarizes our findings on the income distribution of US neighborhoods. We find that a substantial degree of income mixing characterizes the great majority of urban neighborhoods where mix of incomes is the outcome of market forces. In particular, in 1993 over two thirds of the micro neighborhoods included at least one household (out of 10) with an income of 30% of HAMFI (the poorest one sixth of the sample); over half of all neighborhoods included at least one household with an income of 150% of HAMFI (the richest one fifth of the sample3). Section 4 reviews our conclusions and briefly discusses policy implications.