The Rise of the Bipolar Neighborhood

In its darker past, urban planning was sometimes used as a tool to rigidly segregate different groups into distinct neighborhoods. During the last half-century, however, the tenets of urban planning have clearly evolved toward the goal of communities providing for a diversity of income (and, often, ethnic) groups in residence (Atkinson & Kintrea, 2000; Cole & Goodchild, 2001; Sarkissian, 1976). These tenets have been reflected in numerous tangible ways, from the Garden Cities/New Towns movement (Minton, 2002), to the concerns over the wholesale replacement of original residents by “gentrifiers” (Freeman, 2006; Smith, 1996), to being a key ingredient of “sustainable communities” (Tunstall, 2003). Indeed, the current AICP Code of Ethics and Professional Conduct states that “we shall seek . . . to promote . . . economic integration” (American Planning Association, 2005). This article investigates the degree to which metropolitan neighborhoods have moved toward or away from this goal since 1970, why change may have occurred, what consequences may transpire, and what implications this holds for planners. In particular, our study employs the traditional nominal entropy (H) index, the newly developed ordinal entropy (E) index, and their ratio (E/H) to probe changes in the distribution of incomes within neighborhoods. We analyze census tract data for 1970–2000 in the 100 largest metropolitan areas Problem: Although planners aim to provide for income diversity in the communities they serve, too little is known about how income distributions in metropolitan neighborhoods are changing. Purpose: We investigate whether neighborhood income diversity has increased since 1970 by examining neighborhoods in the 100 largest U.S. metropolitan areas. Methods: We analyze neighborhoods in the 100 largest U.S. metropolitan areas from 1970 to 2000 using a combination of nominal (H) and ordinal (E) entropy indices. We focus on neighborhoods we call bipolar, (where E/H > 1), in which very lowand very high-income groups predominate. We investigate these with tract-level statistics and by using a counterfactual. Results and conclusions: We find a dramatic increase in the number and incidence of these bipolars since 1970. Compared to other neighborhoods, we find that, on average, bipolars have significantly greater shares of very high-income families, racial diversity, shares of middle-aged persons, and shares of renters. We use a counterfactual to reveal that much of the growth in bipolars over the last three decades has been fueled by income distributions at the metropolitan scale becoming more bimodal, with fewer middle-income families. Gentrification appears to explain only a minor share of growth in bipolars. Takeaway for practice: Metropolitan census tracts with pronounced bimodal income distributions have become more common since 1970. This appears to reflect changing metropolitan income distributions more than spatial rearrangement, although planning policies may be responsible in some instances. Whether residence in bipolar neighborhoods will benefit very low-income households by reducing stereotyping and expanding social opportunities is unclear, but such places should be monitored.