Security versus Status? - CiteSeerX

10.1177/0739456X04270127 Sanchez Security versus et al. Status

ARTICLE

Security versus Status?

A First Look at the Census’s Gated Community Data

Thomas W. Sanchez, Robert E. Lang, & Dawn M. Dhavale

T

he term gated communities for most people conjures up images of exclusive developments with fancy homes and equally fancy lifestyles. At the gates stand guards who screen all nonresidents or the uninvited. Much of the popular and academic literature on gated communities promotes this view (see Garreau 1991; Blakely and Snyder 1997; Lang and Danielsen 1997; Stark 1998; Low 2003). These authors also focus on how some gated communities closely control the lives of residents, including extreme examples such as limiting the number of guests allowed to parties or the types of vehicles that one can park in a driveway. Gated communities are also easy targets for social critics who point to their walls as the physical manifestation of a long-standing exclusionary impulse among rich people to shut out the less fortunate (including a big chunk of the middle class) (Guterson 1992, 1993). Such criticism extends to popular culture, including an X-Files episode several years ago where a monster eats those who fail to follow the homeowners’ association rules or a Twilight Zone episode where unruly teenagers are turned into fertilizer. While much of the attention has focused on the demographic characteristics and geographic distribution of upscale gated communities, little attention has been devoted to other dimensions of enclosed communities represented by low-income, renter households. Recent data released by the U.S. Census Bureau as part of the 2001 American Housing Survey (AHS), shows that low-income renters are actually more likely to live in walled or gated communities compared to affluent homeowners. Because class and race are correlated in the United States, the owner and renter distinction translates into a separation of high-income from low-income and Whites from nonWhites. While affluent White homeowners in gated communities have been extensively profiled, the gated, low-income, non-white renters have not. We suspect these two worlds reflect a divide between gated communities, one the result of status versus one motivated by concern for security. Security is arguably the motivating force behind the trends in gated communities. The “forting up” of households, especially affluent households, has been widely reported. Walls and gates are physical barriers to prevent intrusion by an undesirable element associated with crime, drugs, vandalism, and disregard for public or private property. Walls and gates defend and protect residential space from these activities and, Journal of Planning Education and Research 24:281-291 DOI: 10.1177/0739456X04270127 © 2005 Association of Collegiate Schools of Planning

Abstract For most people, the term gated communities conjures up images of exclusive developments with fancy homes and equally fancy lifestyles. Much of the popular and academic literature on gated communities promotes this view. Yet the common perception of gated communities as privileged enclaves turns out to be only partly correct based on our analysis of the first ever census survey of these places. There are gated communities composed of mostly White homeowners with high incomes that have a secure main entry—the kind of classic gated community in the public mind. But there are also gated communities that are inhabited by minority renters with moderate incomes. We expected that this dichotomy reflects a divide between gated communities, one based on status versus one motivated by concern for security. Using the 2001 American Housing Survey (AHS), we attempted to explain the differences between gated homeowners, nongated homeowners, gated renters, and nongated renter households. Keywords: gated communities; housing; American Housing Survey Thomas W. Sanchez is currently an associate professor of urban affairs and planning at Virginia Tech’s Alexandria Center and a fellow with the Metropolitan Institute. He holds a Ph.D. in city planning from the Georgia Institute of Technology. His research has been in the areas of transportation, land use, residential location behavior, and questions of social equity in planning, with research being sponsored

281

282

Sanchez et al.

in so doing, also create security both socially and economically. Protecting residential property values becomes nearly as important as physical safety.1 An accurate portrait of who lives in gated communities and why they choose these places is important to the public policy debate over community development. In some markets, a significant share of affordable rental housing is gated (for an example, see Figure 1). Many moderate-income households pick gated communities for reasons that remain not entirely clear. The easy guess is that they have safety concerns, but an underlying motivation may be that the gates signal a higher level of amenity and status than comparable nongated developments. If residents select gated apartments primarily for safety reasons, that could signal a retreat from community engagement and result in what Blakely Figure 1. Multifamily development surrounded by fences. and Snyder (1997) describe as “enclaves of fear” (p. 99). In such places, the relatively less disadvantaged turn their backs on neighbors by hiding behind gates. But if the primary selling point for downscale gated communities is that they are packaged with pools and other amenities then the motivation may be a lifestyle choice rather than fear. This is an important distinction for public policy makers because the first motivation signals a potential loss of community while the latter indicates a consumer choice that may not represents a desire to exclude others. Therefore the answer to our question, “security versus status?” speaks to the larger meaning of gated communities. Should planners and policy makers concern themselves that the rise of gated communities signals a fundamental shift in the nature of by the National Science Foundation, neighborhood life? Perhaps they should, but only after a careful examination of why U.S. Environmental Protection Agency, Fannie Mae Foundation, U.S. Department such places attract residents. We find that the new AHS data offers a starting point in of Housing and Urban Development, answering basic questions about gated communities that until now have mostly been Brookings Institution, and state departspeculated on. The AHS for the first time provides baseline data, which in itself ments of transportation. includes the counterintuitive finding that gated communities are more prevalent Robert E. Lang is the founding director of among downscale renter than upscale owner markets. The data also provides insight the Metropolitan Institute at Virginia (albeit limited) into what causes people to choose gated communities and, correTech in Alexandria, Virginia, and an associate professor in Virginia Tech’s School of spondingly, what implications these motivations have for those concerned with Planning and International Affairs. Dr. understanding neighborhood change. Lang is coeditor of the new scholarly publication Opolis: An International Journal of Suburban and Metropolitan Studies. He also serves as associate editor for the journal Housing Policy Debate, and book review editor for the Journal of the American Planning Association. He is currently working on a book for Brookings Institution Press entitled Boomburbs: The Rise of America’s Accidental Cities. Dawn M. Dhavale is a doctoral student in urban affairs and planning at Virginia Tech in Alexandria and a researcher at the Metropolitan Institute at Virginia Tech. She has conducted research on topics such as micropolitan areas, census-designated places, and suburban growth trends. Her dissertation research focuses on metropolitan development. She has a masters in urban and regional planning from Virginia Tech.

䉴 Background Residential “fortification” is an underlying theme of previous research on gated communities. Typical gated communities are entirely surrounded by physical walls with gates, and sometimes restrict entry with guards or other means of access control. Enclave communities are removed from the main road and built with a cohesive feeling but lack unbroken physical barriers (Langdon 1994). Enclosed, or security zone, communities (Blakely and Snyder 1997) are created when once-public streets are blocked off to create a private community. Enclosed or enclave communities have also been retrofit with perimeter walls with a guardhouse or gate after the initial construction. Another permutation is the “blockhome”—an individual home with security devices well beyond the average security system (Flusty 1997). This article examines gated communities, both with and without access control. According to the literature, households more frequently choose gated housing for security and prestige. The aims of security and prestige appear vastly different, but they

Security versus Status 䉳

are not mutually exclusive and neither aim is principally associated with household income level. These two desires do not appear to have much in common, but they do have common themes, including physical control, psychological benefits, and protection from crime. The ability to control the physical environment is one of a gated community’s most powerful abilities. By reducing and screening through traffic, prestige communities create an exclusionary status, and security communities reduce the perceived risk of intruders and criminals. A gate or a guard creates an image of either a privileged land or perimeter that is difficult to breach (Flusty 1997). Gated communities can also control the social environment both directly and indirectly. Indirectly, house prices and location control community entry, effectively restricting certain groups or segregating the community by class and race (Judd 1995; Blakely and Snyder 1997). Affluent Americans have employed a number of devices to control housing practices in the past, particularly restrictive covenants— agreements concerning individual ability to purchase property that was conveyed with the house—and large lot, minimum house size, and minimum frontage zoning (Higley 1995). Unlike real estate covenants of the past, gated community restrictions are legal and are rarely challenged for being discriminatory. By excluding unwanted persons or activities, gates are an effective way to protect property values. Almost all gated communities have homeowners’ associations to protect property values by controlling their neighborhood environment (Fischel 2000; Lang 2003). Blakely and Snyder (1997) characterize gated or walled communities as a means for moderate to upper income homeowners to separate themselves economically, socially, and psychically from low-income, minority households. Walls and gates prevent intrusions associated with crime, drugs, vandalism, and disregard for public or private property. In so doing, walls and gates create social and economic security and project a desired image. Gated communities offer psychological benefits for some households. The gate is an affirmation of social status to affluent households and provides a way to keep out the unknown. The gate is a reminder of the unpredictability of the world outside the gates (Guterson 1992). Gates symbolically remove a community from the outside world and all of its troubles (Judd 1995). Although residents may actually be no safer inside the gates, residents in gated communities report that they feel less safe when outside of their communities (Low 2001). Gated communities are marketed as safer, but are they really? The few studies on this topic are mixed. Newman (1980) reported a decline in crime on private streets as compared to public streets in St. Louis, while others suggest that there is no difference (Wilson-Doenges 2000; Blakely and

283

Snyder 1997). With the emphasis on security, it is important to know, Is the prevalence of crime rising, or is the actual incidence magnified? Crime rates are going down, but fear of crime is rising (Colvard 1997). With intense media coverage and contradicting statistics, the perception of crime can increase regardless of actual changes in crime rates (Glassner 1999). For instance, polls show that Americans are increasingly afraid of being the victim of crimes in both public and private spaces (Lofland 1998). Current information specific to location and situation is rarely available. Research in this regard suggests that targeted information can reduce inflated perceptions of crime (Stone 1996). Gated communities may deter crime by using some of the crime prevention through environmental design (CPTED) concepts. It is generally believed that crime reduction can occur through group behavior and environmental changes (Newman 1972). Gating is “target hardening,” making the potential target much more difficult to access (Crowe 1991). However, this access restriction simultaneously affects the legitimate users of the environment. Engaging the community is paramount, and enclosing or restricting areas makes residents feel responsible for the area, drawing them together through shared responsibility—by having more “eyes on the street” (Jacobs 1961). Yet gates may counterintuitively make communities less safe. Once the responsibility for vigilance has passed to a hired security force or an electronic code system, neighbors may be less likely to keep watch on suspicious people or activities in the neighborhood (Lang and Danielsen 1997; Blakely and Snyder 1997). Media accounts and academic research has referred to lowincome gated communities as security zones. Many have been retrofit on significant decline (National Public Radio 1995; Financial Times 1997; Low 1997; Davis 1990). One of the few studies to address low-income gated communities looked at the difference between low- and high-income gated and nongated communities in terms of safety and “sense of community” (Wilson-Doenges 2000). For high-income households, gated communities, contrary to commonly held perceptions, had a lower sense of community.2 In addition, there was no reported difference in the sense of community between low-income gated and nongated communities. While crime rates differ little between gated and nongated communities in low- and high-income neighborhoods, low-income households did not report any difference in perception of safety or community cohesion. The high-income groups did not report differences by day, but at night, the high-income gated community residents reported feeling safer than their nongated counterparts. The high-income gated residents also felt that their neighborhood was safer than other neighborhoods, even though there were no differences in actual crime rates

284

between neighborhoods. The low-income communities did not report any comparative safety differences. Gated communities are paradoxical. They are marketed as close-knit communities while effectively reducing the interaction between neighbors. Internally, they are hotbeds of volunteerism, yet they may reduce civic engagement in the larger community (Lang and Danielsen 1997). There is no clear consensus on whether they reduce crime or protect property values. Much of the evidence on these topics is anecdotal and contradictory. Developers and planners also contend that gates can increase a feeling of community (Blakely and Snyder 1997), despite research that suggests otherwise (WilsonDoenges 2000; Merry 1993). Related to this is the gated residents’ disengagement from the local community and involvement only in issues relevant to their immediate community (Lang and Danielsen 1997). Closing off previously public streets and building new gated communities can restrict public community space to those that live in the gated community. The newly private spaces are sometimes “militarized” with cameras and other security devices (Davis 1990, 223-63; Judd 1995). New communities are built with private roads, swimming pools, meeting areas, and other amenities that are not open for public use. Many walled or gated neighborhoods are selected because of these amenities and for privacy, however, not the feeling of community (Merry 1993). Some research suggests that the desire for privacy increases with household income. Consequently, by choosing higher levels of protection, residents also forgo typical levels of neighborhood social interaction. Social control is no longer exercised by direct face-to-face contact, but instead by authority figures, usually a private security force employed by the association. The role of social control was historically more personal, with neighborhood forums compared to today’s guards and association board pronouncements (Baumgartner 1988). This is likely a reflection of the self-selection of residents who value privacy and wish to restrict certain social contact, as well as dissolution of community. Some do not care about their neighbor’s opinions so much as they want to avoid being fined or face lawsuits for not abiding by the association rules. Sending a security guard to complain about a neighbor’s barking dog avoids the risk of a negative confrontation.

䉴 Research Approach While much of the discussion surrounding gated communities focuses on perceived civic and social costs of exclusive and homogenous neighborhoods, little research has at-

Sanchez et al.

tempted to distinguish household types that reside in these types of communities. One explanation for this is that until the 2001 AHS, no nationwide household data existed that would allow such distinctions. Using survey responses from questions that characterize household types, unit types, community types, residential location, mobility status, along with indicators of residential preferences are analyzed to determine whether security or status are effective predictors of who lives behind walls and who does not. Drawing from the previous research, we hypothesize that indicators of status will play a significant role in distinguishing households living in gated communities. In addition, we expect that indicators of safety may be more important for low-income renters, especially those in urban core areas.

䉴 Methods The 2001 AHS added forty new questions ranging from types of home financing, country of origin for household members, and community attributes of residential locations. For the first time, the national sample included questions to distinguish gated communities, two of which are

• Is your community surrounded by walls or fences preventing access by persons other than residents?

• Does access to your community require a special entry sys-

tem such as entry codes, key cards, or security guard approval?

Using responses from these two questions, we examine the characteristics of households that live in “walled” or “access controlled” communities. This analysis segments these households by tenure status, race, ethnicity, socioeconomic status, and household composition relative to walled and access controlled communities. The responses to these questions are first used to profile the households by neighborhood types and also to detect the primary differences between homeowners in gated communities, homeowners in nongated communities, renters in gated communities, and renters in nongated communities. If gates or walls surrounding communities provide safety, security, or other psychological benefits, it should be reflected in the responses provided about neighborhood quality, preference, and satisfaction. The attractiveness and benefit of particular amenities should also be expressed in the household responses for reasons that the current residential unit and neighborhood were selected. Along with descriptive statistics, a discriminant analysis was performed to distinguish the four household types on the basis of thirty-four selected variables from the AHS. Discriminant

Security versus Status 䉳

analysis produces linear combinations of the independent variables to predict group (tenure and community type) membership. We hypothesized that the resulting discriminant functions would provide useful insights about the distinguishing characteristics between gated and nongated households as well as by ownership status. We anticipated that the factors related to status, such as unit size, unit and neighborhood design, community facilities, school quality, and construction quality, would be more strongly associated with gated homeowners. On the other hand, we expect that factors related to safety, such as levels of neighborhood crime, abandoned buildings, barred windows, broken windows, and neighborhood cleanliness, would be associated with gated renters.

285

Table 1. Top 10 metropolitan areas. Top 10 Metropolitan Areasa Atlanta Boston Chicago Dallas Detroit Houston Los Angeles New York Philadelphia Washington, DC

% Walled

% Access Controlled

7.4 3.5 5.3 17.8 2.3 26.7 18.2 5.2 2.0 4.3

5.5 0.6 1.3 13.4 1.2 21.9 11.7 1.7 0.8 2.6

a. Alphabetic listing based on 2000 population.

䉴 Results Of the 119,116,517 housing units represented by the AHS, 106,406,951 were occupied year-round, with 7,058,427 (5.9 percent) households reporting that they lived in communities surrounded by walls or fences and 4,013,665 (3.4 percent) households reporting that access to their communities was controlled by some means.3 Only respondents who said they lived in a walled or fenced community answered the survey question about controlled access, which means that nearly 60 percent of the walled or gated communities also had controlled entries (i.e., gates). The percentages of walled or controlled-access communities vary by region of the country, with households in the West having a higher likelihood of living in walled communities (11.1 percent), followed by the South (6.8 percent), the Northeast (3.1 percent), and the Midwest (2.1 percent). One explanation for the regional distribution of walled and gated communities is that they are more prevalent in new construction, with significant amounts of new residential development in the West and South. The regional concentration of walled and gated households is also reflected at the metropolitan scale (see Table 1).

Owner versus Renter Household Characteristics Contrary to the notion that primarily affluent homeowners live in gated communities, the results of the AHS survey show that renters are nearly 2.5 times more likely to live in walled or fenced communities and over 3 times as likely to have controlled entries. These renters include households in public housing projects, which often have walled and gated design elements. The survey data also show that owners and renters have significantly different demographic profiles, with owners

more likely to be white compared to renters (86.4 percent compared to 67.1 percent), to have higher incomes ($73,548 compared to $35,831), to have older heads of households (fifty-two years old compared to forty-two years old), and to have slightly larger households (2.7 persons compared to 2.3 persons). Tables 2, 3, and 4 provide a comparison of household characteristics relative to tenure status. Table 5 shows these household characteristics across the income groups (quartiles), which indicate that higher income households are actually less likely to live in walled or gated communities compared to lower income households. Higher income households are also predominantly white, younger, and have more persons than the lower income groups.

Community Types Examining community types across the dimensions of race, ethnicity, and tenure status shows that Hispanics, whether homeowners or renters, are more likely to live in walled or controlled entry communities than whites or blacks. In contrast, black homeowners are the least likely group to live in either type of community, with the rates for black renters falling between those of white and Hispanic households (see Table 6). The question arises about why there are smaller proportions of blacks living in walled or controlled access communities. Is it a function of the housing market not providing adequate choice for this segment of the population (affluent blacks) or simply an aversion to walled or gated communities? To test this, a subsample that included only responses from the Atlanta, Baltimore, Birmingham, Detroit, Memphis, and Washington, DC, metropolitan areas was analyzed. While blacks constitute 12.5 percent of the nation’s households, in these metropolitan areas, they represent nearly 30 percent. It would seem that affluent blacks would have a

286

Sanchez et al.

Table 2. Demographic characteristics of walled and unwalled developments. Households

% Walled

All Walled/fenced Not walled/fenced Access controlled Access not controlled

5.9 100.0 0.0 99.7 98.6

% Access Controlled 3.4 56.7 0.0 100.0 0.0

% White

% Black

% Hispanic

Mean Agea

Mean Household Income

Mean Household Size

80.3 70.6 81.0 70.8 70.3

12.5 14.6 12.3 14.8 14.4

8.2 15.7 8.8 13.8 18.0

48.76 47.00 48.89 45.70 48.60

61,481 60,562 61,566 66,343 52,749

2.55 2.29 2.57 2.12 2.52

a. Mean age of household head.

Table 3. Demographic characteristics of walled and unwalled owners. Owners

% Walled

% Access Controlled

% White

% Black

% Hispanic

Mean Agea

Mean Household Income

Mean Household Size

All Walled/fenced Not walled/fenced Access controlled Access not controlled

4.3 100.0 0.0 100.0 98.9

2.1 49.2 0.0 100.0 0.0

86.4 84.4 86.6 87.2 82.0

8.7 6.3 8.8 4.0 8.4

6.5 10.6 6.4 7.1 13.9

52.14 54.52 52.04 54.95 54.13

73,548 86,731 72,998 105,467 68,773

2.66 2.41 2.67 2.22 2.59

a. Mean age of household head.

Table 4. Demographic characteristics of walled and unwalled renters. Owners

% Walled

% Access Controlled

% White

% Black

% Hispanic

Mean Agea

Mean Household Income

Mean Household Size

All Walled/fenced Not walled/fenced Access controlled Access not controlled

11.6 100.0 0.0 99.6 98.4

7.3 62.6 0.0 100.0 0.0

67.1 59.7 68.1 60.6 57.8

20.5 21.3 20.4 21.6 20.8

14.9 19.7 14.3 17.9 22.4

41.58 41.01 41.65 39.94 42.66

35,831 39,758 35,333 42,003 35,527

2.33 2.20 2.35 2.05 2.44

a. Mean age of household head.

Table 5. Income and demographic characteristics for walled and controlled access residents. Income Quartile 1. 2. 3. 4.

% Walled

% Access Controlled

% Owner

% White

% Black

% Hispanic

Mean Agea

Mean Household Size

7.5 7.1 5.8 6.1

4.1 4.1 3.1 3.8

51.0 60.0 73.6 88.3

73.0 78.4 83.4 86.6

19.4 14.1 9.5 6.7

10.7 11.2 9.1 5.8

55.05 47.93 45.18 46.50

1.93 2.39 2.81 3.12

< $20,000 $20,000-$40,000 $40,001-$71,768 > $71,768

a. Mean age of household head.

Table 6. Tenure and race/ethnicity by community type. % Fenced or Walled Communities

Owner Renter

White

Black

4.2 10.3

3.1 12.0

% Controlled Entry Communities a

Hispanic 7.0 15.2

a

White

Black

Hispanic

2.1 6.6

1.0 7.6

2.3 8.7

a. Includes heads of households reporting race/ethnicity as white Hispanic and black Hispanic.

Security versus Status 䉳

287

Table 7. Descriptive statistics. Gated Homeowner

Nongated Homeowner

Gated Renter

Marital status 0.602 Household income 87,794 # of persons in household 2.407 White 0.850 Black 0.060 Asian 0.057 Other (race) 0.034 Hispanic 0.105 Male 0.607 Central city status 0.266 Northeast region 0.099 Midwest region 0.082 South region 0.421 West region 0.397 People in neighborhood are bothersome 0.023 Liked unit because of room layout/design 0.097 Liked unit because of exterior appearance 0.051 Liked unit because of construction quality 0.058 Liked unit because of size 0.070 Liked unit because of yard/trees/view 0.041 Moved to obtain higher quality unit 0.033 Chose neighborhood for its looks/design 0.087 Chose neighborhood because of good schools 0.027 Neighborhood has neighborhood crime 0.121 Windows covered with metal bars 0.051 Windows broken 0.021 Factories/other industry within a half block 0.020 Mobile homes within a half block of unit 0.147 Bodies of water within a half block of unit 0.276 Neighborhood police protection satisfactory 0.941 Community recreational facilities avail 0.606 Abandoned/vandalized buildings within a half block 0.016 Trash/junk in streets/properties in a half block 0.044 Buildings with bars on windows within a half block 0.073 Weighted N 2,959,717

0.646 73,172 2.672 0.868 0.087 0.023 0.022 0.063 0.630 0.221 0.184 0.259 0.366 0.192 0.034 0.062 0.034 0.031 0.046 0.035 0.029 0.054 0.028 0.122 0.034 0.037 0.025 0.138 0.171 0.925 0.317

0.315 39,735 2.221 0.599 0.208 0.081 0.112 0.202 0.534 0.536 0.094 0.086 0.421 0.399 0.049 0.148 0.064 0.041 0.127 0.049 0.050 0.128 0.061 0.245 0.077 0.030 0.063 0.056 0.141 0.945 0.558

0.300 35,461 2.362 0.686 0.201 0.038 0.075 0.145 0.480 0.432 0.227 0.214 0.325 0.234 0.049 0.111 0.046 0.030 0.123 0.048 0.061 0.095 0.059 0.216 0.065 0.065 0.060 0.087 0.123 0.922 0.335

0.536 61,819 2.561 0.806 0.123 0.030 0.041 0.092 0.584 0.293 0.190 0.235 0.358 0.217 0.039 0.080 0.039 0.031 0.071 0.039 0.039 0.069 0.038 0.153 0.045 0.044 0.036 0.121 0.160 0.926 0.340

0.037

0.053

0.075

0.048

0.069

0.127

0.145

0.092

0.057 65,567,497

0.116 3,682,851

0.116 28,072,568

0.076 100,282,633

Variable a

Nongated Renter

Total

a. “Marital Status” is a dichotomous variable where married = 1 and unmarried = 0.

greater opportunity to choose walled or gated communities in those metros where they make up a substantial proportion of their respective populations. The survey results indicate, however, that black homeowners in the six selected metros are actually less likely to live in walled or gated communities compared to national averages. Only 1.6 percent of black homeowners lived in walled or fenced communities, with only 0.3 percent living in communities with controlled entries. On the other hand, 11.8 percent of black renters lived in walled or fenced communities, and 8.4 percent lived in communities with controlled entries in these metros—virtually the same rates as for all renters across the United States.

Table 7 shows the descriptive statistics for the thirty-four selected variables and the four classes of household types. As expected, there are several clear differences among tenure status, as well as gated status. Compared to renters, homeowners are more likely to be married, affluent, white, and residing in suburban locations. Homeowners also reported having fewer instances of bothersome neighbors, less crime, and fewer abandoned buildings, litter, and barred windows compared to renters. Homeowners were less likely to be near industrial land uses but more likely to be near mobile home sites. Land-use zoning and the locational preferences of homeowners explain being further from factories and industrial sites. Mobile homes

288

are also more likely to be located away from central cities at the urban fringe or outlying areas. It is surprising to note that homeowners (especially gated homeowners) did not report school quality, unit design, exterior appearance, or quality of construction as significantly different locational factors compared to the other household types. In fact, for gated homeowners, good schools, unit size, and unit appearance were not as important as they were for gated renters, nor was obtaining a higher quality unit a major factor in homeowner’s reasons for moving to their current location. Overall, gated homeowners separate themselves from poor quality residential environments by choosing neighborhoods that have fewer physical signs of deterioration like broken windows, trash, and barred windows. While renters reported that they perceived higher levels of crime where they lived compared to homeowners, there was virtually no difference reported between gated and nongated homeowners in this regard. In addition, gated households reported that neighborhood police protection (used to measure perceptions of safety) was satisfactory slightly more often than did nongated households, whether they were homeowners or renters. While it is difficult to directly observe the degree to which status plays a role in location patterns, we expected that the physical appearance of the housing or neighborhood would emerge as a consistent set of indirect measures. The survey results suggest that residence size, design, and appearance were often as important for renters as they were for homeowners. And while it was expected that gated households would feel safer with respect to crime, there was no significant difference reported between gated and nongated homeowners, with gated renters reporting higher levels of concern for crime compared to nongated renters. There was also virtually no difference in the perception of police protection across tenure or gated status.

Discriminant Analysis The discriminant analysis included the thirty-four independent variables shown in Table 7, with household type (gated homeowners, nongated homeowners, gated renters, and nongated renters) as the criterion variable. The independent variables fell into three general categories that represented demographic characteristics and safety considerations. The model was able to correctly classify approximately 71 percent of more than 100 million households observations. In the first function (gated homeowners), six demographic variables had the strongest explanatory power (see Table 8). These results serve to reinforce the stereotype of gated homeowners as being married, affluent, white, and living in the suburbs. It

Sanchez et al.

was anticipated that the variables associated with status would play a more substantial role, especially for this group. The variables related to the size of the unit, room layout/design, neighborhood appearance, landscaping, view, and school quality were negatively correlated with the likelihood of being a homeowner in a gated community, contrary to what was expected. In terms of the perception of safety variables, gated homeowners were less likely to be concerned about crime and had less neighborhood trash or junk and fewer buildings with bars on the windows near them. The second and third functions do not exhibit patterns of discriminant loadings as clearly as the first function.

䉴 Discussion This article presents the first analysis of AHS data that included survey responses about households in walled and gated communities. One notable finding was the prevalence of low-income, racial minority renters reporting that they lived in walled or gated residential developments. This represents a departure from even the most recent literature profiling the characteristics of gated communities (see Low 2003). With these findings, we expected that residential preferences would play a significant role in distinguishing patterns of residential location type. In particular, we hypothesized that variables associated with security and status would differentiate gated homeowners from nongated homeowners, gated renters, and nongated renters. This was not, however, supported by the descriptive statistics or discriminant analysis presented here. Despite the results of the discriminant analysis that show demographic variables explaining a high degree of variation among household types compared to expressed preferences for residential unit or community characteristics associated with perceived security or social status, obviously, demographic characteristics are endogenous and highly correlated with residential opportunity and preferences. Additional research is needed to untangle these dimensions to better understand the motivations of households that choose walled or gated residential environments. The results of the analysis also prompt a reexamination of what it means to be “walled” or “fenced” and how it affects residential space. Walled or fenced communities are usually depicted as a function of affluence or exclusivity, whereas the opposite may actually be more accurate. Mobile home parks and public housing projects are typically enclosed spaces, but these barriers do not effectively provide increased safety or any measure of status. The implications of this research will likely extend beyond the demographic differences among gated and nongated

Security versus Status 䉳

289

Table 8. Structure matrix from discriminant analysis. Function Variable

1 a

Marital status Central city Household income White Hispanic Black Liked unit because of size Neighborhood has neighborhood crime Trash/junk in streets/properties in a half block Buildings with bars on windows within a half block # of persons in household Factories/other industry within a half block Liked unit because of room layout/design Chose neighborhood for its looks/design Abandoned/vandalized buildings within a half block Windows covered with metal bars Chose neighborhood because of good schools Liked unit because of yard/trees/view Community recreational facilities available Midwest region Asian Liked unit because of exterior appearance Neighborhood police protection satisfactory Bodies of water with a half block of unit Male Northeast region Windows broken Mobile homes with a half block of unit South region Moved to obtain higher quality unit Liked unit because of construction quality People in neighborhood bothersome

b

0.621 –0.435b 0.421b 0.408b –0.282b –0.264b –0.263b –0.237b –0.234b –0.198b 0.191b –0.184b –0.177b –0.158b –0.147b –0.143b –0.141b –0.067b –0.074 0.120 –0.110 –0.068 0.002 0.115 0.256 –0.052 –0.107 0.129 0.052 –0.140 0.000 –0.069

2

3

–0.139 –0.076 –0.241 0.057 –0.129 0.096 0.005 0.019 0.132 0.008 0.093 0.047 –0.085 –0.091 0.142 –0.050 0.043 –0.010 –0.566b 0.373b –0.209b –0.074b –0.071b –0.220 –0.088 0.309 0.164 –0.004 –0.188 0.063 –0.100 0.052

–0.337 –0.240 0.145 0.326 –0.123 –0.134 0.181 –0.141 –0.017 0.114 –0.162 –0.079 –0.016 0.033 0.040 0.018 –0.002 0.064 0.156 –0.144 –0.140 –0.013 –0.035 0.435b b –0.361 b 0.325 b 0.245 b 0.198 b –0.194 b 0.182 b 0.139 b –0.087

Note: Pooled within-groups correlations between discriminating variables and standardized canonical discriminant functions. Variables ordered by absolute size of correlation within function. a. “Marital Status” is a dichotomous variable where married = 1 and unmarried = 0. b. Largest absolute correlation between each variable and any discriminant function.

households over time. While issues of housing affordability, satisfaction, and choice are implicit to examining residential location behavior, increased personal safety and security concerns may also be a more prominent factor given the events of September 11, 2001, and threats of terrorist attacks. Additional data from subsequent AHS surveys and other survey research efforts can be used to test whether safety concerns will be reflected in residential development and patterns of household demand. The first journalistic account based on these preliminary findings revealed some interesting case accounts from residents of rental gated communities. Haya El Nasser (2002), in a cover story for USA Today, found that safety was just one dimension of why people chose to live in these places. In conversa-

tions with residents of rental gated communities in suburban New Orleans, El Nasser reported the following: Gated access is “a safety issue not so much from criminal elements,” says Henry Shane, president of an architectural firm in New Orleans that has developed 25 gated apartment complexes, including some for moderate-income renters. “It limits the amount of people who can come in.” That’s what drew waitress Gina Rojas to one of Shane’s developments. Harper’s Ferry is not fancy. But the iron gates at all the entrances give it a tidy look and discourage outsiders from parking in spaces reserved for tenants or littering the grounds. (p. A2)

In their fieldwork for an upcoming book on fast-growing suburbs, Lang and Lefurgy (forthcomimg) found that a key

290

Sanchez et al.

Authors’ Note: The authors wish to thank Patrick Simmons, Diane Zahm, Karen Danielsen, and two anonymous reviewers for their helpful comments on previous drafts.

䉴 Notes

Figure 2. Gated rental under construction in Moreno Valley, California.

reason why so many gated rental communities appear in the suburbs is that these places are acceptable to middle-income homeowners. Gated rental communities blend seamlessly into places dominated by master-planned community development. This community acceptance of affordable gated renters resulted in the city of Moreno Valley, California, requiring that all new multifamily housing be gated (see, e.g., Figure 2). In addition to further statistical work, researchers should now focus on ethnographic and descriptive analyses of rental gated communities, especially those in the suburbs. These places have so far gone largely unstudied, despite the fact that the majority of gated community residents live in rental units. Our own preliminary read of rental gated communities (and El Nasser’s work) hints at a myriad of motivational forces behind a preference for these types of units. We suspect that a major factor may be that gated rental developments signify “middle-class” respectability that distinguishes these places from typical suburban garden apartments. As noted above, it may even be that the gates are bundled with a host of other amenities (such as pools) that give these developments a higher quality of life than the nongated apartment complexes. An ethnographic account of rental gated communities is needed to uncover the complicated forces that lie behind the decision to live in these places. Finally, because 2001 was the first year that questions on gated communities were included on the AHS, there is no way to determine whether the increasing popularity of these places is tending toward a more racially integrated mix of residents. The 2001 AHS provides the first national glimpse at these patterns, and subsequent national and metropolitan surveys will be useful for detecting these trends.

1. We realize that the two concerns are not mutually exclusive. Upscale gated communities typically sell security, but the walls are often more a marketing tool to signify high status (Blakely and Snyder 1997; Lang and Danielsen 1997). Downscale gated communities offer security as a more pragmatic response to high crime in comparable nongated neighborhoods (Blakely and Snyder 1997). 2. Wilson-Doenges (2000) refers to “sense of community” as the “set of networks among people who share interactions” (p. 598). 3. This does not include high-rise structures with locked or guarded entries.

䉴 References Baumgartner, Mary Pat. 1988. The moral order of a suburb. Oxford: Oxford University Press. Blakely, Edward J., and Mary Gail Snyder. 1997. Fortress America: Gated communities in the United States. Washington, DC: Brookings Institution Press and Lincoln Institute of Land Policy. Colvard, Karen. 1997. Crime is down? Don’t confuse us with the facts. HFG Review 2 (1): 19-26. Crowe, Timothy. 1991. Crime prevention through environmental design. Stoneham, MA: Butterworth-Heinemann. Davis, Mike. 1990. City of quartz: Excavating the future in Los Angeles. London: Verso. El Nasser, Haya. 2002. Gated communities are not just for the wealthy. USA Today, December 16, pp. A1-2. Financial Times. 1997. The birth of enclave man. September 20. Fischel, William. 2000. The home voter hypothesis: How home values influence local government taxation, school finance and land use politics. Hanover, NH: Department of Economics, Dartmouth College. Flusty, Stephen. 1997. Building paranoia. In Architecture of fear, edited by N. Ellin, 47-60. New York: Princeton Architectural Press. Garreau, Joel. 1991. Edge city: Life on the new frontier. New York: Doubleday. Glassner, Barry. 1999. The culture of fear. New York: Basic Books. Guterson, David. 1992. No place like home: On the manicured streets of a master planned community. Harper’s Magazine, November, pp. 55-64. . 1993. Home, safe home. Utne Reader, March-April. Higley, Stephen R. 1995. Privilege, power and place: The geography of the American upper class. Lanham, MD: Rowman & Littlefield. Jacobs, Jane. 1961. The death and life of great American cities. New York: Random House.

Security versus Status 䉳

Judd, Dennis R. 1995. The rise of new walled cities. In Spatial practices, edited by H. Ligget and D. C. Perry, 144-65. Thousand Oaks, CA: Sage. Lang, Robert E. 2003. From virtue to value: Redefining the role of homeownership in America. Keynote talk by Robert Lang at the Western Knight Center for Specialized Journalism, the Annenberg School for Communication, University of Southern California. March 31. http://www.mi.vt.edu/ ConferenceSchedule/USC%20Web%20Posting.doc. Lang, Robert E., and Karen A. Danielsen. 1997. Gated communities in America: Walling out the world? Housing Policy Debate 8 (4): 867-77. Lang, Robert E. and Jennifer Lefurgy. Forthcoming. Boomburbs: The rise of America’s accidental cities. Washington, DC: Brookings Institution Press. Langdon, Philip. 1994. A better place to live. Amherst: University of Massachusetts Press. Lofland, Lyn. 1998. The public realm: Exploring the city’s quintessential social territory. New York: Aldine de Gruyter. Low, Setha M. 1997. Urban fear: Building fortress America. City and Society, (Annual Review): 52-72.

291

. 2001. The edge and the center: Gated communities and the discourse of urban fear. American Anthropologist 103 (1): 45-58. . 2003. Behind the gates: Life, security and the pursuit of happiness in fortress America. New York: Routledge. Merry, Sally Engle. 1993. Mending walls and building fences: Constructing the private neighborhood. Journal of Legal Pluralism 33:71-90. Newman, Oscar. 1972. Defensible space: New York: Macmillan. . 1980. Community of interest. Garden City, NY: Anchor Press. National Public Radio Transcript. 1995. L.A. gated communities. In Rights and the common good, edited by Amitai Etzioni. New York: St. Martin’s Press. Stark, Andrew. 1998. America, the gated? Wilson Quarterly 22 (1): 58-79. Stone, Christopher. 1996. Crime and the city. In Breaking away: The future of cities, edited by C. Stone, 83-103. New York: The Twentieth Century Fund Press. Wilson-Doenges, Georjeanna. 2000. An exploration of sense of community and fear of crime in gated communities. Environment and Behavior 32 (5): 597-611.