Keywords

1 Introduction

The mainstreaming of lesbian and gay culture through the adoption of same-gender marriage may be changing the nature of lesbian and gay enclaves (Ghaziani 2014). Stories in the press (James 2017) note this changing nature of gay neighborhoods as bars, clubs and bookstores that cater primarily to a gay and lesbian clientele close. These closings are evidence of neighborhood change and displacement. Urban planners usually use Census data to measure neighborhood change. Yet the Census, even in 2020, does not explicitly ask about sexual orientation or gender identity. This Chapter develops an alternative way of looking at LGBTQ neighborhood change using Census data. This allows for the development of a baseline to allow further assessment of LGBTQ neighborhood changes. The resulting analysis gives some context to the generational rise and relative decline of the gayborhood.

Measuring US neighborhood change relies heavily on Census data in the United States. Since the 1940 Census, urban areas of the United States have been divided into census tracts (Snow 2011). For the last 80 years, urban scholars analyzed changes in population and housing within neighborhoods to the degree that census questions and definitions defined variables of interest. These variables include age, sex, race, marital status, housing characteristics, household composition, and work status. With each Census, definitions of variables would be modified to reflect necessary changes in proposed outcomes and to reflect changes in social conceptions of subgrouping in work and residential life (Alonso and Starr 1987). The lack of Census questions on LGBTQ variables such as gender and sexual orientation has meant that analyses of LGBTQ neighborhoods and places rely on a combination of other information sources to determine their location and degree of concentration (Forsyth 2011). Early analyses of gay neighborhoods relied on ethnographic stories of LGBTQ community members building neighborhood institutions (Castells 1983). Other methods for analysis included identifying concentrations of LGBTQ institutions and organizations such as bars, businesses and non-profits who were willing to list themselves in LGBTQ guides (Harry 1974; Levine 1979; Wolfe 1992). Historical work has had to rely on a combination of oral history and archival material (Chauncey 1995). Without a national gay rights law, LGBTQ people risk their own livelihoods by being out and counted as part of a community. Such readings of history have had to rely on the ability to read code—implicit expressions of queerness identifiable between the lines (Sedgwick 1990). Throughout most of the twentieth century, the ethics of compulsory heterosexuality and the closet meant that much evidence of variance may have been destroyed to protect reputations (Rich 1993). Thus, it was an exciting advance in urban studies in the late 1990s when the Census began to put together a series of tables on same-sex partnered households (Black et al. 2000). These data might allow a more accurate accounting of residential LGBTQ neighborhoods.

Researchers used this dataset. Most famously, Richard Florida’s Creative Class (2002) theory relied on the Census Bureau’s concentration of same-sex partnered households as a measure of regional tolerance (Florida and Gates 2003). At the same time that the Creative Class theory was gaining respectability, the Census Bureau was involved in a process of reassigning lesbian and gay partnered responses as heterosexual responses when constructing these data sets (Simmons and O’Connell 2003). The reassigning of responses is gay removal. Evidence about changes in gay neighborhoods based solely on this dataset must account for changes in how the Census Bureau constructs the dataset (MRFHS 2014). While 25 years of Census data on same-sex partnered households exists, the 2020 Census does not ask the questions about sexual orientation and gender necessary to develop a fuller understanding of LGBTQ communities and neighborhoods (Doan 2016; FIWG 2016b; Edgar et al. 2018; Wang 2018).

The answer to this lack of data is to do a queer reading of the census. This requires reading the structured silences within Census data (see Frisch 2002). Census data and government survey data has been structured around questions of citizenship, representation, and distribution (Alonso and Starr 1987). These structures reflect heteronormative ideals of marriage, households and the nuclear family. Therefore, a queer reading of the Census asks—how would a queer person answer the Census questions? How and where does my household and the households of LGBTQ friends and acquaintances show up in the data categories collected? Asking these questions raises issues about intersectionality, the closet, gender expression, passing, and what it means to live in and around LGBTQ communities. Such a reading allows for identifying clusters of LGBTQ individuals and positions LGBTQ neighborhoods in contrast to the question of non-heteronormative neighborhoods.

This analysis starts with a quick review of heteronormality and its social enforcement by urban development processes and planning. Census questions and the resulting data, must be considered within this context. The next section presents the problems of Census data on LGBTQ communities. The most accurate way of assessing LGBTQ neighborhoods would be to add questions about sexual orientation and gender identity. The earliest we might get an accurate census count of LGBTQ communities is 2030. Without these data, the queer reading of the Census proceeds by asking, “how might a person with a non-heteronormative life answer the Census?” Indicators such as sex ratio, and marital status may then identify neighborhoods outside heteronormative expectations. The analysis proceeds by illustrating how a queer reading of the Census might work with examples of four probable “queer” neighborhoods—Midtown in Kansas City; Midtown in Atlanta; the West Village in New York City and Center City Philadelphia. Such a reading provides evidence of a decline in queerness—indicating areas of possible displacement by the end of the study period. This chapter concludes that full recognition in the public realm requires counting.

2 Heteronormativity and Urban Development

While different sexual orientations and gender identities have been around forever, the meanings attached to the categories are products of the modern era (D’Emilio and Freedman 2012). Homosexuality and heterosexuality are terms coined by sexologists of the late nineteenth century and only really attained their modern meanings around the turn of the century. Heterosexuality as a term arose to encompass often unspoken assumptions about normal society. Katz (1995) identifies three basic components of heteronormality: that marriage can only be between one man and one woman, that a nuclear family of heterosexual parents is the expected and best site for raising children, and that it is the only site where expression of romantic and sexual pleasure should be allowed. These assumptions about heteronormality still surround us every day—who comprises an average family? What gets shown on TV, and what is suitable for kids to see?

Frisch (2002) argues that modern methods of urban planning arose at the same time as these categories of sexual orientation identity. Urban planning acts to build and promote heterosexual spaces and places on purpose. The rise of suburbs made of single-family houses is a part of this heterosexist project as apartment buildings are seen to threaten sexual and gender norms. The development of LGBTQ neighborhoods post-Stonewall must be viewed within the context of systematic power expressed through societal pressure and physical environments favoring heteronormality as well as the direct powerful legal forms of discrimination. These threats lessened as more people came-out and joined social movements demanding their rights. Some see the LGBTQ rights movement as one of the most successful twentieth-century movements for social change (Sullivan 2005; Lakey 2018).

Coming out, a successful organizing tactic of LGBTQ politics, is an act of performance (Butler 1993). When LGBTQ folks choose to be out and publicly express gender and sexual identity, they act against these cultural norms. LGBTQ rights have been achieved because of the millions of people choosing to contradict the assumptions of heteronormality. Yet, the “closet” functions as a “double bind” (Halperin 1995). It is a strategy to protect yourself from the physical violence of hate crimes, and from discrimination at your job. In this sense it gives you agency. It is also the strategy of received and perceived oppression that restricts what you do (Sedgwick 1990). The closeted safe choice is to stay silent. As Foucault (1978) notes, these silences have patterns. Finding these structures of individual responses to political, social, and cultural oppression then requires both an understanding of possible inconsistencies in response, as well as a view of how heterosexuality is empowered through legal and extralegal means. Even same-sex marriage may be seen through this lens. A same-sex marriage of course overturns the hierarchy of sex roles within marriage. Two lesbians together raising kids disorients the assumptions of heterosexuality in that masculine and feminine roles in parenting may be performed by someone of the same gender. The arguments for religious freedom in regard to LGBTQ folk are all about defending the primacy and natural existence of different sex roles (Alliance Defending Freedom 2019). These cases however hover around the discovery that a possible client, customer or patient is LGBTQ. Information management still matters.

The normalization of LGBTQ status then requires tolerance, recognition, and eventually acceptance of LGBTQ status. Tolerance might be seen through the early stages of LGBTQ enclave formation (see Forsyth 2011). Recognition requires some acceptance of standing as a citizen as evidenced by the Supreme Court rulings in the Lawrence v. Texas (2003), US v. Windsor (2013) and the Obergefell v. Hodges (2015) decisions. Increasing acceptance may then lead to a more dispersed residential pattern (Ghaziani 2014). Without LGBTQ places then do we still need LGBTQ spaces (Nusser and Anacker 2013, 2015)? If gay neighborhoods are in decline, do we then need to take action to protect and plan for LGBTQ space? Planning for LGBTQ neighborhoods and communities requires information about individuals and households comprising the community. In the United States, the Census provides initial local data that informs planning analysis.

3 The Census, Heteronormativity, and LGBTQ Populations

US Census questions change every decade to reflect changing notions of American citizenship (Anderson 2015). Furthermore, the Bureau of the Census runs and coordinates other social surveys such as the Current Population Survey in order to collect information necessary for further implementation of government policies and programs (Alonso and Starr 1987). The 1990 Census added questions about unmarried partners in households (Simmons and O’Connell 2003), however, as the data was being collected, the assumption was being made that unmarried couples had to be of a different sex. Responses that were from people in same-sex partnerships had their responses changed by the Bureau to being a response of a different sex. This was during a health crisis when accurate data about gay men would have saved lives—yet the Census Bureau was actively removing lesbians and gays responses from the Census. With the passage of the Defense of Marriage Act in 1996, these acts of gay removal became the policy of the Bureau. A technical note from 2003 stated:

Same-sex spouse responses were flagged as invalid to comply with the 1996 Federal Defense of Marriage Act (H.R. 3396) passed by the 104th Congress. This act instructs all federal agencies only to recognize opposite-sex marriages for the purposes of enacting any agency programs. In order for Census Bureau data to be consistent with this act and the data requirements of other federal agencies, same-sex spouse responses were invalidated. The legislation defines marriage and spouse as follows:

… interpretation of the various administrative bureaus and agencies of the United States, … the word ‘spouse’ refers only to a person of the opposite sex who is a husband or wife.—Simmons and O’Connell (2003)

This policy interfered with how same-sex partner datasets could be developed. If the partner answering the Census question used the word “spouse” it must be invalid. The Census Bureau was acting as an enforcer of heteronormality. This decision of course had an impact on the datasets used by Florida and Gates in developing their tolerance index (2003). In the 2000 census, if someone listed their partner as partner they were counted as being in a same-sex unmarried couple. If they said they had a spouse—their response was reclassified as “straight.” The tolerance index then measured the degree that LGBTQ folks in same-sex partnerships used the term “partner” in answering the census.

How much did this process of gay removal impact the same-sex partner database? Later work on the 2010 Census data revealed the degree of gay removal and sex classification errors in the Census data. Using “uncorrected data” from the full-count, same-sex partnered households who used the term “partner” accounted for 0.32% of all households in the year 2000 and 0.47% of households in the 2010 census (O’Connell and Feliz 2011: 5). Same-sex households that used the word “spouse” to describe their partnership accounted for 0.24% of all households in 2000 and 0.30 in 2010. According to counts produced by Census procedures “spouse” households accounted for 43% of these households in 2000 and 41% of households in 2010. Yet, Gates and Steinberger’s (2009) work found that only 16% of unmarried partners would answer the census describing their partner as spouse. It turns out that the over count of “spousal” responses in the Census was due to a Census procedure accounting for non-respondents to the Census. The Census uses the probability that a particular name aligns with a particular sex to assign a sex classification to people in non-respondent households (O’Connell and Feliz 2011). Errors due to this sex-assigning name processing led to a 28% over count of same-sex partners in 2010 (O’Connell and Feliz 2011: 23). This sex assignment procedure may also produce a “misgendering.”

The 2010 Census occurred at a moment when the Defense of Marriage Act was still the Federal law, and only Massachusetts, Connecticut, Iowa, Vermont, New Hampshire and DC had legalized same-sex marriage. More than half the states had some sort of state law or constitutional ban on same-sex marriage. The 2010 Census definition of the family maintained that a family required one or more people living in the same housing unit who are related to the householder by “blood, marriage, and/or adoption.” These criteria have been consistent for 80 years (Pemberton 2015). In the 2010 Census many same-sex partnered couples with kids were only considered to be a family due to their having kids whereas an opposite-sex couple who was married without kids would be considered a family. Once again, heteronormality is the underlying deciding factor in how the datasets are put together between families and households.

Federal agencies reacted to the problems with organizing data around same-sex partners and unmarried couples and the challenges that increasing recognition of same-sex partnered relationships made to heteronormal assumptions about marriage and families. In 2010, during the Obama administration, the Office of Management and Budget organized the Interagency Working Group on Measuring Relationships in Federal Household Surveys to examine issues related to collecting information about household formation. After reviewing 55 surveys done by various Federal agencies, this task force found a series of measurement issues that could lead to inconsistent results:

  1. 1.

    variation in response categories (for example, more categories and/or different category wording);

  2. 2.

    inconsistent measurement of relationship to child;

  3. 3.

    infrequent measurement of interrelationship of all household members;

  4. 4.

    inconsistent measurement of cohabitation;

  5. 5.

    infrequent measurement of sex for all household members; and

  6. 6.

    inconsistent inclusion of State of current residence and State where married—(MRFHS 2014: ii).

The task group also suggest various ways to improve data gathering:

First, question wording should incorporate sex‐neutral language wherever possible;

Second, Federal surveys should continue working to collect information on intimate relationships other than opposite‐sex marriage:

Third, Federal agencies should review their current use of editing and processing procedures with respect to sex, relationship, and marital status;

Finally, results from tests and discussions should be widely shared, not only among Federal agencies but also with other interested parties such as academic and policy research organizations.—MRFHS (2014: iii)

The task group also remarked that the United States v. Windsor (2013) case overturning the Defense of Marriage Act would have an impact on how survey data are processed (MRFHS 2014: 3). The work of this task group then set the stage for potentially asking about sexual orientation and gender identity in the 2020 Census. While reporting on questions for the Census and the Current Population Survey one of these research groups found no “significant issues that would make collecting SOGI (sexual orientation and gender identity) information in the CPS infeasible” (Edgar et al. 2018: 4). Draft lists of questions for the 2020 Census including sexual orientation were initially proposed in 2018 only to be censored by Trump administration officials (Wang 2018). Only recently have Federal agencies begun the work of understanding the implications of asking questions about gender identity (FIWG 2016a; Holzberg et al. 2018).

Table 3.1 sorts out the various national estimates of the number of same-sex partnered households in the United States within the context of all partnered households whether married or not. Over the last two decades the number of same-sex partnered households has more than doubled going from 0.6% of households in 2000 to 1.33% in 2019. Perhaps this is evidence of a rise in homonormativity (Bell and Binnie 2004). At the same time, we still do not have population-based numbers for sexual orientation and gender identity. We do not know how many un-partnered people consider themselves LGBTQ. We do not know how transgender and/or genderqueer folks would answer the “sex”-based questions on the survey; and we do not know how many opposite-sex partnered households are made up of LGBTQ folks living in what looks like a heteronormative household in census terms. Sadly, we must now wait until 2030 to get population-wide results.

Table 3.1 Same and opposite sex coupled households in the United States by unmarried status
Full size table

4 A Queer Reading of the Census

The previous discussion showed why the distribution of the partnered data may or may not estimate the relative degree of concentration of LGBTQ folks in a particular neighborhood. Looking back over time can we find variables where residents answer Census questions in patterns that show that they are not living in a typical heteronormal fashion? While Census questions were written with the presumption of heteronormativity, what if we look for queer patterns instead? When faced with a Census questionnaire, how might have a queer person have answered it? For example, gay partners in New York City often kept their separate apartments. Space was always running short in New York and if you had access to a rent-stabilized apartment you had to be sure about the relationship before giving it up. While some housing protections for LGBTQ New Yorkers have existed since the early 1990s, same-sex marriage makes it easier to guarantee security. In Kansas City, with more LGBTQ repression, queer men and women may have been in opposite-gender marriages in the past. Single LGBTQ folks may have lived in a relatively closeted situation within their kinship networks in one space and lived out their LGBTQ lives by going out and dating in other spaces and places. How will these situations show up in neighborhood census data? A queer reading of the census chooses multiple variables that might reflect these situations.

A concentration of LGBTQ folks would then show up as a concentration of people without the characteristics of heteronormality. From 1960 to 2000 this would mean people of child-rearing age who are not involved in what was counted as marriage at the time—only opposite sex people with a marriage license from the state. Generally, people between the ages of 25 and 54 who are single and/or divorced are not conforming to the societal norm of marriage and procreation during their child-rearing years. The Census has been reporting marital status of people age 15 and older and this variable includes counts of single and divorced people by sex. LGBTQ neighborhoods will therefore have concentrations of these people. With heteronormality, there would be a relative evenness in the sex ratio—the ratio of men to women within the same age cohort. As lesbian and gay social connections develop between members of the same-sex, a lesbian neighborhood may show a higher number of women than men while a gay male neighborhood will have the opposite ratio showing higher numbers of men than women. It is important to note that other socio-spatial forces and institutions may create sex ratio imbalances. For example, mass incarceration leads to minority neighborhoods with a higher proportion of women. Local military bases may lead to neighborhoods with higher proportions of men. Furthermore, the census definition of family also reinforces heteronormality by requiring ties by blood and/or marriage. LGBTQ neighborhoods then will have higher proportions of non-family households.

5 Testing the Variables in Four Neighborhoods

The three variables: the proportion of non-family households, marital status, and sex ratio by age cohort were analyzed for four neighborhoods in four different cities: the West Village in New York City, Center City in Philadelphia, Midtown Atlanta, and Midtown Kansas City. Maps detailing the Census Tracts (US Census Bureau 2020) encompassing the study areas are shown in Fig. 3.1 for New York, Fig. 3.2 for Philadelphia, Fig. 3.3 for Atlanta, and Fig. 3.4 for Kansas City. All of these neighborhoods were chosen because they housed LGBTQ bars in the early 1990s as listed in the Damron guide (Damron Co. 1993, see Knopp and Brown 2020 for an analysis of the impact of these guides). The West Village is the site of the Stonewall Riots in 1969. Center City Philadelphia includes both Washington Square that Jane Jacobs (1961) called a “pervert park” as well as Rittenhouse Square and Camac Street that were notorious as gay meeting places. Midtown Atlanta was chosen to give an idea of how these variables might work within the South. Midtown in Kansas City was chosen in order to capture places where there was a concentration of bars in the past as well as an effort to develop a lesbian community in the 1970s and early 1980s. All of these areas have also faced development and gentrification pressures since at least 1980. With these geographies selected, the question becomes, will the selected variables show the expected concentrations over time? This would be the period from 1960 to 2000 which captures the emergence of the LGBTQ social movements. The second question then becomes, if the variables work as a measure, might changes in these variables reveal increasing and decreasing levels of concentration? This second question gets at the issue of gentrification and displacement. LGBTQ displacement would lead to lower levels of these variables appearing sometime in the years 1990–2015.

Fig. 3.1
figure 1

(Source Map by author)

The West Village in New York

Full size image
Fig. 3.2
figure 2

(Source Map by author)

Center City Philadelphia

Full size image
Fig. 3.3
figure 3

(Source Map by author)

Midtown Atlanta

Full size image
Fig. 3.4
figure 4

(Source Map by author)

Midtown Kansas City

Full size image

6 Non-family Household Results

The proportion of all households comprised of non-family households for each of the selected neighborhoods compared to the result for the United States as a whole is shown in Table 3.2. At the national level, the percent of non-family households has increased in every period. Kansas City, Philadelphia and New York all had three times the number of non-family households in 1960, while Atlanta had more than twice the level of non-family households. By 1970 all four neighborhoods were made up of a majority of non-family households and by the year 2000, seven out of ten households were non-family households in these four areas. Yet, note that the percent of non-family households in the West Village declines in the 1980s and the percent increase in Midtown Kansas City and the West Village lag behind the national change from 1980 to the year 2000. Midtown Atlanta had the highest overall growth in this variable, while Center City, Philadelphia had the least growth of the four neighborhoods, yet Center City and the West Village started at a significantly higher level.

Table 3.2 Non-family households as a share of total households
Full size table

7 Never Married by Sex Results

The percent of men and women who have never married (single people) is shown in Tables 3.3a, b. These results show that there is a difference by sex. Interestingly, there is a relatively consistent higher percent (5–7%) of men who have never married than women over the almost sixty-year period for the United States. Midtown Kansas City had a lower percentage of men never married than the US average in 1960, but had the highest total growth in net percentage over the time period (See Table 3.3a). By 1990 all four neighborhoods had 25% more male singles than the national average. Midtown Atlanta and Midtown Kansas City had much more growth in this factor than Center City and the West Village; but all four neighborhoods had much higher growth than the US overall (Table 3.3a). All four neighborhoods had higher levels of single women than the US at the start of the study period. The West Village started much higher than the other neighborhoods in 1960 and Midtown Kansas City had slightly more single women than single men (Table 3.3b). The net percent growth for single men surpasses the percentage for women for the US and all four neighborhoods for the period, with the neighborhoods having a much higher net difference. Center City Philadelphia single women total percentage growth comes closest to the male percentage growth. By the year 2000, single women make up around 50% of all women in the four neighborhoods. Finally, there was a decline in the percent of men who were single in the West Village in the 1980s and a decline in the percent of women who were single in Center City, Philadelphia in the 1990s. Apart from Kansas City, the other neighborhoods decline in numbers of single men from 2000 to 2010. More recently this decline continues in Midtown Atlanta and Center City Philadelphia up to 2015.

Table 3.3a Share of men, age 15 and over never married by neighborhood 1960–2000
Full size table
Table 3.3b Share of women age 15 and over never married by neighborhood 1960–2000
Full size table

8 Divorced by Sex Results

The percent of women and men over the age of 15 who listed their marital status as divorced are shown below in Tables 3.3c, d. In this question, the Census privileges being married; no matter if it is the second, third, or fourth marriage. All four neighborhoods have higher levels of divorced men in 1960 than the United States as a whole and the levels increase in all four neighborhoods to 1980. After 1980, the percent of divorced men drops in all four neighborhoods even though the percent is rising in the United States. By the year 2000, the level in Center City is less than the US average as shown in Table 3.3c. The percent of divorced men and women in these neighborhoods is higher in Midtown Atlanta and Midtown Kansas City than it is in Center City Philadelphia or the West Village. If you add the percent of men in the four neighborhoods who are either divorced or single, it accounts for 66% of all men in the four neighborhoods by the years 1990 and 2000. The totals for women are generally less, running between 55 and 63%. The net percent growth in divorced men and women lagged behind the growth in the nation for three of the four neighborhoods with Kansas City being the exception. Note that men have a higher rate of never marrying, but women have a higher rate of being divorced as shown in Tables 3.3a–d. By 2010 all four neighborhoods lag in the net growth of divorced men and women compared to the nation.

Table 3.3c Share of men age 15 and older, divorced, by neighborhood, 1960–2000
Full size table
Table 3.3d Share of women age 15 and older, divorced, by neighborhood, 1960–2000
Full size table

9 Sex Ratio of Age 25–54 Cohort Results

The sex ratio measures the number of men in the age cohort relative to the number of women. Within a heteronormative model this ratio should be close to one as children are always assumed to have a married mother and father living together. A sex ratio of greater than one indicates more men than women in that geography and a sex ratio less than one indicates more women than men. The results for the neighborhoods are shown in Table 3.4a. First, note that the US ratio has gone from 0.96 to 1.01 from 1960 to 2000. In 1960, Midtown Atlanta and Midtown Kansas City had more women than men while Center City and the West Village were about average. The ratio rises to hit the peak in 1980 in Midtown Atlanta, Center City and the West Village and 1990 in Midtown Kansas City. The ratio increases are quite sharp for Midtown Atlanta and Midtown Kansas City and much more balanced in Center City and the West Village. Both Center City and the West Village had significant lesbian and gay male communities in 1960; did these communities concentrate in particular census tracts within these neighborhoods? Variation of sex ratios within each neighborhood’s individual tracts is shown in Table 3.4b. This analysis shows a much, much higher concentration in specific census tracts in the West Village and Center City in 1960. The highest ratios were in Center City and the West Village in 1960. The patterns of highs and lows were different in each place. Midtown Atlanta started as a woman dominated area in 1960, and only with 1970 did a tract have significantly more men. By the peak of sex ratio difference in 1980, all Midtown Atlanta tracts had more men than women, a pattern that continues to a lesser extent to this day (Table 3.4b). In Midtown Kansas City, a tract had the highest number of women per men in 1960 of all four neighborhoods. The peak in the tract-based sex ratio difference was in 1990 in Kansas City, In Center City Philadelphia, there have been both male-dominated tracts and female dominated tracts consistently over the study period. While the Center City sex ratio difference dropped from 1960 to 1980, it rose again by 1990, dropped by 2000 and rose by 2010. Finally, the West Village had one of the greatest differences in sex ratio in 1960 only to drop to the least difference after the year 2000.

Table 3.4a Sex ratio of age 25–54 cohort by neighborhood compared to US 1960–2000
Full size table
Table 3.4b Sex ratio of age 25–54 cohort variation extremes across neighborhood census tracts
Full size table

10 Discussion

These measures capture related ways of examining the concentration of folks possibly leading LGBTQ lives in neighborhoods thought to be LGBTQ friendly in the last 55 years. Each of the tables shows how queer lives in the neighborhoods substantially differed from the United States. The direction of the indicators toward concentration in these factors creates evidence of the rise of LGBTQ community formation with the baby-boomer generation forming what we now know as gayborhoods. The neighborhoods trend together up until 1990 or 2000 and then the trends on the studied factors become more variable across the four neighborhoods. The nuances in the data trends are also interesting. Midtown Kansas City and Midtown Atlanta both had higher proportions of divorced men than Center City Philadelphia and the West Village in New York. This possibly reflects stronger enforcement of “compulsory heterosexuality” in the urban South and the urban Midwest than on the East Coast especially in the 1970s and 1980s.

The reduction in concentration post 1980 may reflect a reduction in the value of proximity. However, the specificity also smacks of displacement. The 1980s and 1990s were the HIV plague years for gay men with hundreds of deaths in these neighborhoods. Sarah Schulman (2013) calls the losses of this time “the gentrification of the mind.” AIDS deaths significantly altered the gender make up of these neighborhoods. Each of the neighborhoods had significant new developments that displaced LGBTQ institutions. Starting in the mid-1980s, Midtown Atlanta along Peachtree was transformed from a low-rise district to a mid-rise and high-rise district (Doan and Higgins 2011). After 1990, Midtown Kansas City replaced a neighborhood with gay bars and clubs with Midtown Marketplace housing a Home Depot and a Costco. Center City Philadelphia created the University of the Arts south of City Hall replacing LGBTQ bars. Finally, the West Village underwent multiple transformations as the West Side Highway was transformed into a boulevard and the piers became parks. New residential developments and luxury lofts replaced artist studios and LGBTQ clubs. LGBTQ folks get displaced by these gentrification processes. The peaks in non-family households in these neighborhoods around 1990 may also reflect the lesbian and gay baby boom beginning with lesbian couples in the 1990s and gay male couples after the year 2000 (Gates 2013; Gurrentz and Valerio 2019). These households would reduce the number of non-family households, while having a lesser impact on the sex ratio.

11 Comparison to Same-Sex Unmarried Partner Data

The changes in the non-family household variable raises the question about how same-sex partnered households are currently counted in the American Community Survey (ACS). Table 3.1 shows that by 2019 a majority of same-sex partnered households were married in the latest Current Population Survey data. How does the flawed same-sex unmarried partner data in 2000, corrected ACS data for 2010 (2008–2012 5 Year data) and corrected ACS data for 2015 (2013–2017 5 Year data) compare to the results of these measures? The same-sex partner data confirm that there are relatively high concentrations of same-sex partners in these four neighborhoods as shown in Table 3.5. The concentration is three to ten times more in these neighborhoods than the overall rate for the nation as a whole. These households still make up a small fraction of the households in these neighborhoods. Yet, the queer reading of the census analysis showed that these neighborhoods have a majority of households that reject heteronormality in one form or another.

Table 3.5 Unmarried same-sex partners as a share of all households in the four neighborhoods
Full size table

12 Conclusion

This analysis took the first step toward a queer reading of the census. Exploratory factor analysis might be used to deepen the analysis by uncovering other variables that capture the rise of LGBTQ communities. Such an analysis may be able to distinguish underlying factors that also lead to changes in levels of the selected variables such as the sex ratio. For example, during the period analyzed in this study, mass incarceration removed a significant generation of minority men from their neighborhoods. This removal would show up in terms of low sex ratios for these neighborhoods. Other variables could be used. The rate of detached single-family homes may be a marker for heteronormality as fixed by the Euclid V. Ambler decision (Frisch 2002). Smart and Klein’s (2013) findings suggest mass transit use might also be an indicator.

Even at their peak, LGBTQ neighborhoods were really just enclaves (see Marcuse 1997). There were always other folk in these neighborhoods. This analysis shows an increase in concentration and then a flattening out and/or decline in the measures. The declines in concentrations show evidence of queer displacement.

With increased social acceptance, along with increased legal rights due to three positive Supreme Court cases: Lawrence V. Texas (2003), US v. Windsor (2013), and Obergefell v. Hodges (2015), LGBTQ folk might not need their own neighborhoods and enclaves to the same degree as in the past (Kelly et al. 2014). Intersectional LGBTQ communities may not use space and place in the same ways as stereotypical white gay men (Irazábal and Huerta 2016). Worry about social acceptance for sexual orientation and gender identity may now play less of a role in choosing a place to live (Ghaziani 2014). Queer Millennials may not seek the same qualities in neighborhoods as LGBTQ Boomers. After the 2016 election, there is evidence of backlash (Miller 2019). This backlash has been accompanied by the rise of anti-LGBTQ spaces—health providers, pharmacists, wedding cake bakers, and florists who claim that their religious beliefs are being violated by equally serving LGBTQ folks (Melling 2018; Green 2019a, b). The “violation” arises in opposition to someone who is truthful, public, and out about their LGBTQ identity which results in further LGBTQ displacement.

While these methods will be necessary to assess LGBTQ neighborhoods in the past, it is unsettling that sexual orientation and gender identity questions will not be on the 2020 Census. This omission reveals that the Census will still treat gender as “sex.” This failure means the “queer reading of the Census” will still be necessary into the next decade. Good answers to these questions will be hard to get as long as people can be fired for their answers. Getting the data is a significant part of recognition. It also would provide key data that can be used to develop services and inclusive urban plans (Forsyth 2011). The lack of asking the questions, reinforces the notion that speaking about sexual orientation and gender identity is somehow offensive. It indirectly provides support to business establishments and service providers whose discriminatory actions toward LGBTQ people are being increasingly protected (Green 2019b). Full recognition in government services and social surveys would be an important step toward establishing and protecting LGBTQ rights.