ABSTRACT

It is a virtual truism in the field of urban studies that hot, sunny places always grow, or do they? The deep and rich literature on the American Sunbelt tells the story of the attractive powers of these warm environs-where the introduction of air conditioning made them ideal for jobs, workers, and retirees. A 1926 cover story in Liberty magazine suggested that there were no brakes on the Florida growth engine (Cave 2009). Cheap land and abundant amenities made Arizona, Texas, New Mexico, South Carolina, and other states desirable in the second half of the twentieth century beyond any booster’s wildest dreams. The new alignment of power, money, and people away from the Rustbelt to the Sunbelt has been viewed by many scholars since the 1970s as largely permanent (see Perry and Watkins 1977; Weinstein and Firestine 1978; Bernard and Bradley 1983; Sawers and Tabb 1984; Schulman 1994; Pack 2005). Except for the occasional boom-bust cycles, the Sunbelt has been viewed as a place of perpetual prosperity and limitless growth. In this chapter, I introduce empirical evidence to refute that idea and to begin to fill in the conceptual framework introduced in Chapter 5. Sunbelt cities are presently in a state of economic and population decline: they are shrinking. More importantly than demographic change, many of these declining Sunbelt cities are physically shrinking, presenting an unprecedented opportunity to reshape and retrofit urban neighborhoods in a positive way. This late twentieth-century thinking about the Sunbelt has been shaped primarily by interpretations of past data. Bowman and Pagano’s (2004) Terra Incognita: Vacant land and urban strategies report on a survey and Census data analysis that was influential at the time and serves as a useful introduction to my own findings. The study involved a survey mailed to city officials in U.S. cities with populations of 50,000 or more from 1997 to 1998 (531 were sent, 186 responded) and a second survey mailed to 81 with follow-up questions. The survey results showed a massive amount of vacant land in Southern (n=23) and Western (n=30) cities, 20,011 acres and 10,349 acres, respectively. With Midwestern and Northeastern cities reporting 5,903 and 5,004 acres, on average. The survey also asked local officials to report the number of abandoned buildings in their cities (only 20 Southern and 23 Western cities reported data for these questions). Among those who reported data, the Southern cities had an average of 1,632 abandoned buildings and the Western cities had only 93. These numbers are important when contrasted against Northeastern and Midwestern cities that reported on average 4,025 abandoned buildings.1 While there was certainly

as Mobile, Alabama (2,009 abandoned buildings) and Richmond, Virginia (3,000 abandoned buildings) and the growing, new economy cities such as Pembroke Pines, Florida (one abandoned building) and Santa Clara, California (five abandoned buildings). In this new millennium, with the advent of the foreclosure crisis and the Great Recession, it is worth studying the whole of the Sunbelt to understand the impacts and responses for both sides of the coin. But recognizing the historical differences among Sunbelt cities and between Sunbelt cities and other U.S. cities is critical to conducting this analysis. Bowman and Pagano demonstrate this variation with respect to abandonment and through their surveys of vacant land they raise important questions about how undeveloped land fits into local planning strategies (we will return to this point in the case study chapters). In approaching the current shift afoot in the Sunbelt, I relied on two primary data sources that provide demographic and land-use data at the neighborhood level. The first is the well-known and popular U.S. Census, which enumerates population, employment, and housing levels at the city and, in some cases, zip code level. Census data collected on a decennial basis is also available at an even finer geographical level called the census tract (usually smaller than a zip code geography and comprising, on average, a few thousand households). I compiled this data for every Sunbelt city with over 100,000 in population in 2005. In defining the Sunbelt, I adapted boundaries employed by several demographers and geographers to include cities in those states that are located below the 37th Parallel.2 The cities are in the following states:

Due to data errors, seven cities were removed from the analysis, leaving a total of 140 cities in the above 13 Sunbelt states.3 Most (86 percent) of the 140 Sunbelt cities studied were growing rapidly just a few years prior to the housing crash. In total, these cities gained 2,372,033 persons between 2000 and 2005 (474,406 per year, on average). As the housing crash approached in 2006, many of these cities began to change course. But the Sunbelt, as a whole, did continue to grow (albeit at a slower pace). In total, these cities increased in population by 1,138,245 persons between 2006 and 2008. But a closer look reveals a crack in the sustained growth patterns. As housing prices began to sink in 2006, populations also began to dwindle in a select group of cities. Twenty-six Sunbelt cities lost population from 2006 to 2008, including cities in Florida, California, Louisiana, Georgia, Alabama, and Mississippi (see Table 6.1, Figure 6.1). In this group of losing cities, the mean loss was 1,660 people (with a standard deviation of 1,603), while the highest decline was 6,680 persons in Baton Rouge, Louisiana.4