Introduction
Black Mayors Respond to Police-involved Deaths, Protest, and Civil Unrest
The 2020 murder of George Floyd by Minneapolis Police officer Derek Chauvin provoked a broad, impassioned response that has been compared to the country’s reaction to the assassination of Martin Luther King, Jr. (Altman Reference Altman2020). The George Floyd protests also bore striking parallels to demonstrations following the high-profile, police-involved deaths of Michael Brown in 2014 and Freddie Gray in 2015, during the Obama years. Then, reigning Black mayors reacted publicly to police-involved killings and the urban unrest they evoked. However, along with President Obama, they were often characterized as engaging in “racial distancing” and the “politics of respectability” to avoid appearing soft-on-crime or anti-police (Stephens-Dougan Reference Stephens-Dougan2020). By contrast, in 2020 the wholesale reaction of Black mayors seemed different: they were intensely vocal in response to Floyd’s death, using their platforms to decry systemic racism, empathize with protestors, and advocate police reform. In this study, I examine whether Black mayors’ individual and collective responses to George Floyd’s killing were qualitatively different than mayoral reactions to the deaths of Michael Brown and Freddie Gray. Specifically, I examine whether Black mayors rejected the “timeworn” tradition of respectability politics in favor of a different paradigm in 2020. I also consider individual, local, and national political dynamics that may explain this change.
I explore these questions by analyzing the rhetoric of all sitting Black mayors of the 100 largest American cities in response to the killings of Michael Brown, Freddie Gray, and George Floyd. My original dataset includes 232 public statements issued by 31 Black mayors during either the Obama administration (2014–2015) or Trump administration (2020). My analysis of this data partially comports with Harris (Reference Harris2014) and Stephens-Dougan (Reference Stephens-Dougan2020), who have argued that Black political speech during the “Age of Obama” tends to embrace the “politics of respectability.” I confirm that during Era 1, mayors frequently engaged in racial distancing by activating tough rhetoric condemning the behavior of a narrow subset of unruly protestors who damaged property during protests. In 2020, however, I observe a different pattern: Black mayors were significantly less likely to racially distance by engaging in respectability politics when talking about Floyd’s murder and subsequent urban unrest. In doing so, mayors defended their citizens and cities against President Trump’s long-standing tradition of demonizing urban places and people. I discuss the implications of these observations and consider the potential for Black urban pride to be promoted by big city mayors as a counterweight to respectability politics.
Theory
Respectability Politics: A “Timeworn Tradition”
Black leaders, especially former President Barack Obama, have been characterized as routinely cloaking their critiques of American racial dynamics in the politics of respectability. Various prominent Black politicians (Maxine Waters), civil rights leaders (Jesse Jackson), public intellectuals (Ta-Nehisi Coates; Cornel West), and political scientists (Harris 2012; Reference Harris2014; Stephens-Dougan Reference Stephens-Dougan2020) have made this observation. Hallmarks of so-called respectability politics include linking racial uplift to equal rights (Higginbotham Reference Higginbotham1992), the disavowal of Black rage (Smith Reference Smith2014), shunning Blackness and Black culture (Harris Reference Harris2014), the embrace of female-centered bourgeois respectability (Wolcott Reference Wolcott2013), and an emphasis on public safety through respect for law and order (Forman Jr. Reference Forman2017). In brief, the politics of respectability is premised on the idea that Black people must bring their behavior in line with the norms of middle-class, (white) America to enjoy equitable legal protections and prosperity. When Blacks chastise other (especially marginalized) Blacks for improper behavior and self-presentation while urging them to get a job, embrace traditional gender roles, and obey adversarial police, they are engaging in the politics of respectability.
The rhetoric of President Barack Obama and Michelle Obama has been interpreted to contain such elements. As Stephens-Dougan observes in Race to the Bottom, “Obama’s speeches to predominantly Black audiences frequently invoked the themes of absentee fathers and personal responsibility,” (Reference Stephens-Dougan2020, 69). Likewise, although widely praised for her infamous line, “When they go low, we go high,” at the 2016 Democratic convention, Michelle Obama’s “classy” response to Donald Trump’s hectoring treatment of Hillary Clinton could alternatively be viewed as capitulating to bourgeois ideas about feminine docility (Wolcott Reference Wolcott2013). Similarly, Howard (Reference Howard, Julia and Nicol2018) describes the First Lady’s Let’s Move anti-obesity campaign as emblematic of a respectability politics that aimed to distinguish the First (Black) Family from so many Black families who reside in food deserts with limited options for safe, outdoor exercise.
However, the Obamas possess no monopoly on respectability politics: “respectability talk has been employed by many black mayors over the past decades to address declining black educational achievement and criminality” (Harris Reference Harris2014). To illustrate, Harris describes Philadelphia Mayor Michael Nutter’s response to an incident involving a “flash mob” of hundreds of Black youths who overtook the city’s central business district, damaging property and injuring bystanders. In a fiery speech delivered at a prominent Black Protestant church, Mayor Nutter derided youth for, “acting like idiots and fools” in the streets (Harris Reference Harris2014). He then capped off his remarks with a call-and-response instructing Black youths to take down their hoodies, pull up their pants, and “buy a belt!” to which the audience members chanted, “Buy a belt, buy a belt!” (Harris Reference Harris2014).
The imagery of Nutter’s tough talk about rowdy teenagers delivered to Black churchgoers illustrates the crowd-pleasing potential of respectability politics, while the theory of racial distancing helps to explain why such rhetoric persists. First, Black politicians embrace the politics of respectability to appear non-threatening to white voters, especially conservative ones. Experimental research found that likely white voters were more supportive of a hypothetical Black Democratic candidate whose political speech derided African Americans than one whose political messaging was “deracialized” (Stephens-Dougan Reference Stephens-Dougan2020, 96). Additionally, racial conservatives were most supportive of a hypothetical Black Democrat whose campaign speech explicitly activated negative racial stereotypes, urging Black voters to, “quit complaining,” and “get off the couch and get to the voting booth,” (Stephens-Dougan Reference Stephens-Dougan2020, 103). Related to this, the concept of “double-marginalization” theorizes that many Blacks are unoffended when they hear Black politicians condemn the behavior of a subset of the community that is already marginalized, like hoodie-wearing teenagers, absentee fathers, or welfare recipients (Cohen Reference Cohen1999). To the extent that “mainstream” Blacks feel differentiated from marginalized Blacks, they will support racialized tropes, since they do not see themselves as the intended target (Cohen Reference Cohen1999; Stephens-Dougan Reference Stephens-Dougan2020, 18).
High-profile, police-involved deaths of marginalized Black men and boys complicate this dynamic. Respectability politics implies that individuals are at fault when they experience unequal or unfair societal treatment, including when interactions with police turn violent (Starkey Reference Starkey2016; Obasogie and Newman Reference Obasogie and Newman2016). Nonetheless, the racial symbolism of police killings may unite Black Americans from disparate backgrounds in heightened racial consciousness (Stout Reference Stout2020). As such, Black leaders who provide an official response confront a political dilemma. First, respectability politics dictates that Black outrage arising over police violence should be harnessed into well-mannered civil society responses centered around voting and peaceful protest (Smith Reference Smith2014). Then, police killings are therefore potentially catalyzing for Black political mobilization. Finally, racial distancing theory suggests that Black politicians are motivated by political expedience, attuned to voters’ racial attitudes, and accountable to distinct, predictable constituencies. Given this, what constitutes an expedient response to a tragic police-involved killing and its devastating aftermath? In such a racially charged environment, Black leaders face scrutiny from all sides.
For example, President Barack Obama and Baltimore Mayor Stephanie Rawlings-Blake were broadly criticized for their reactions to unrest in Baltimore following the death of Freddie Gray in police custody when, “the city erupted in riots, arson and looting” (Wenger Reference Wenger2015). Conservatives criticized Obama’s remarks for insufficiently supporting law enforcement, while the “Black left” rebuked both Obama and Rawlings-Blake for referring to destructive protestors as “thugs” (Stephens-Dougan Reference Stephens-Dougan2020, 3). With such rhetoric, the President and Mayor presumably hoped to racially distance themselves from unruly demonstrators (Stephens-Dougan Reference Stephens-Dougan2020, 3). Still, Rawlings-Blake later apologized for her verbiage, while Obama did not. What might explain this divergence?
Stephens-Dougan speculates that Rawlings-Blake calculated a higher political risk for using negative racial appeals due to her representing a majority-Black city, while Obama, as the leader of a majority-white nation, perceived a higher political reward for doubling down on racist stereotypes. However, Rawlings-Blake’s apparent reversal needs to be situated in the broader political context.
The mayor’s reference to “thugs” during a press conference on April 27, 2015, came just two days after and in direct response to remarks she issued that were construed as too lenient toward rogue protesters. During a press conference on the evening of April 25, the day the Baltimore riots started, Mayor Rawlings-Blake said, “While we tried to make sure that they [protestors] were protected from the cars and the other things that were going on, we also gave those who wished to destroy space to do that as well,” (Schleifer Reference Schleifer2015). Instantly, video of Baltimore’s Black mayor saying she had given protesters permission to “destroy space” circulated around the media and internet, alongside footage of burning buildings and squad cars, and reports that 15 police officers had been injured during the unrest. So, when Rawlings-Blake stated, “Too many people have spent generations building up this city for it to be destroyed by thugs,” it was during an emergency press conference to walk back her “destroy space” comment, which had been widely construed as anti-police and pro-lawlessness (Chuck 2015).
Racial distancing theory suggests that majority-Black Baltimore would accept their mayor taking a consistently hard line against “thugs,” but Rawlings-Blake’s initial impulse was instead to acknowledge the hurt, frustration, and anger rippling through the community. When this proved to be a political miscalculation as a nationalized audience bore witness to escalating unrest in Baltimore, the mayor retreated, ostensibly hoping to be remembered for condemning “thuggery” instead of defending rioting. Nonetheless, despite being once considered a rising star in Democratic politics, Rawlings-Blake ultimately declined to seek re-election, a decision unambiguously connected to the fallout surrounding Gray’s death (Wenger Reference Wenger2015). Evidently, Baltimore’s mayor could not appease her disparate constituencies during this time of heightened racial tension. Could a racial strategy consistently unmoored from respectability politics have yielded a more favorable political outcome?
Race, Identity, and Urban Politics
Racial distancing is not the only strategy available to Black politicians, especially urban Democrats. For instance, Nelson and Meranto (Reference Nelson and Meranto1977) observe that the great northern migration, which brought large enclaves of southern Blacks to urban centers in the western and midwestern states, naturally provided substantial political capital for Blacks to influence municipal elections. They argue that Black mayors could win by appealing to racial identity with “black power” ideology rooted in, “reconstruction of the black community and its identity, with an emphasis on racial pride and self-esteem,” and “the development of black organizations controlled by blacks, before any coalitions can be formed with whites” (Nelson and Meranto Reference Nelson and Meranto1977, 5). Karnig and Welch (Reference Karnig and Welch1980) also assert the importance of Black political mobilization above white community support in the election of Black mayors. Exemplary of this, Nelson and Meranto regard the 1967 elections of Black mayors in Gary, Indiana and Cleveland, Ohio as cases of growing awareness of Black power among Black city dwellers and suggestive of a “new historical era in northern urban politics” (1977, 4-5).
Relatedly, Hajnal’s (Reference Hajnal2001) study of mayoral elections from 1965 to 1999 found that although white, Black, Latino, and Asian voters all exhibited significant in-group candidate preferences and racial solidarity, these were strongest among the Black community. Additionally, Hajnal broke new ground by showing a relationship between the advent and tenure of Black mayoral leadership in a city and reduced racial resentment and lower perceptions of racial group conflict among moderate and Democratic white voters (Hajnal Reference Hajnal2001). Importantly, Hajnal also found that white urban voting behavior mirrors this pattern: Democrats and moderates will support Black mayors if they are incumbents and/or in cities with legacies of electing Black mayors (Reference Hajnal2007).
The benefits of racial distancing are called into question by such scholarship, along with the observation that during the past decade, roughly one-fifth of the 100 largest American cities have been consistently commanded by overwhelmingly Democratic, Black mayors. Stout (Reference Stout2020) argues that this is explicable due to shifts in the American electorate and party politics since the Obama years, with partisans increasingly sorted by racial attitudes. For instance, opinion polls dating back to 1976 indicate that in 2016 the Democratic party’s base was more racially progressive than during any previous time in US history (Stout Reference Stout2020, 13). Likewise, Reny and Newman (Reference Reny and Newman2020) found that public opinion about the police and racism following the George Floyd protests followed a predictable pattern of partisan polarization. Moreover, since 2016 Black Americans seem to have a stronger sense of “linked fate” (i.e., “what happens generally to black people” affects their lives) and are more supportive of racially targeted government aid and policies (Stout Reference Stout2020, 116-117).
Such developments are plausibly explained by high-profile police shootings and the Black Lives Matter movement making racial discrimination more salient to the masses, and the Democratic party taking note. As evidence of this, Stout shows that news headlines about racial discrimination and Google searches involving racism significantly increased between 2012 and 2016. Similarly, Dunivin et al. (Reference Dunivin, Yaojun Yan, Ince and Rojas2022) found that more social media and news content engaged with “systemic racism” and other elements of BLM’s platform coinciding with the 2020 protests, and that this shift in discourse persisted even after BLM demonstrations subsided. Then, Board et al. (Reference Board, Spry, Nunnally and Sinclair-Chapman2020) found that older, middle-class Blacks were increasingly supportive of BLM, and Fenton and Stephens-Dougan (Reference Fenton and Stephens-Dougan2021) found that Black legislators were equally responsive to constituent emails referencing Black Lives Matter as the long-established NAACP.
Considering these trends, Black mayors during the post-Obama, George Floyd era are logically expected to reject respectability politics more often than their predecessors. I conjecture they will use their bully pulpits to reaffirm Black Americans’ right to self-protection instead. On this, Black liberationists during the civil rights era were explicit about the Black community having a right of self-defense “due to the ineffectiveness of nonviolent tactics in situations where blacks are physically threatened,” (Nelson and Meranto Reference Nelson and Meranto1977, 4). Today, cell phone footage, media coverage, academic scholarship, and numerous online databases have increased the public’s awareness of the ongoing physical threat that police violence poses to the Black community. Considering the ideological flexibility of Black nationalism (Moses Reference Moses1988) and its “enduring” support among Black elites and voters (Brown and Shaw Reference Brown and Shaw2002), Black mayors might resurrect its tenets about self-defense in response to police-involved killings. A response rooted in Black nationalism, then, would promote the right to self-protection while embracing racial pride. Such a stance would also seem to align with the contemporary Black Lives Matter movement.
For instance, some supporters of Black Lives Matter contend that anger and outrage, even when they manifest in looting and rioting, are understandable if not legitimate responses to police killings of unarmed Blacks (Cooper Reference Cooper2014). Exemplary of this, Kerrison et al (Reference Kerrison, Cobbina and Bender2018) found that Black millennials in Baltimore rejected respectability politics in response to Freddie Gray’s death in police custody. As such, scholars influenced by the ideas and activism of Black Lives Matter view the movement as diametrically opposed to respectability politics (Obasogie and Newman Reference Obasogie and Newman2016). In this way, tenets from the Black power and Black Lives Matter movements stand in stark contrast to the rhetoric of respectability and may serve as suitable alternatives. Then, rather than racially distancing, it might behoove Black mayors to embrace Black urban pride as a rebuke to their political opponents.
Practically all cities led by Black mayors are historically associated with sizeable Black constituencies and Democratic party dominance, since such characteristics are key to the emergence of Black political leadership. In turn, conservative politicians routinely attempt to undermine the foundation of Black and/or Democratic urban leadership. As Swanstrom observes, “Running against cities is a time-honored tradition in American politics, with conservative politicians [using] cities as foils to mobilize small town and rural voters. They portray cities as feral places where the welfare state, impersonal bureaucracies, and excessive immigration have corroded traditional moral values” (Swanstrom Reference Swanstrom2017). As such, I hypothesize that the rhetoric of Black mayors in response to national racial issues is mediated by the partisanship and tone of the sitting president.
Following this, perhaps Era 1 mayors embraced Obama’s brand of respectability politics because he did not criticize the intrinsic dysfunction of Ferguson, Baltimore, or their leadership for the unrest that ensured (recall, he blamed “thugs” instead). By contrast, President Trump has a “long history of injecting himself into racially sensitive cases,” (like the 1989 Central Park jogger assault) and an equally lengthy track record of publicly shaming America’s major cities and their leaders as “disasters” of Democratic governance (Colvin and Long Reference Colvin and Long2020; Swanstrom Reference Swanstrom2017). It would seem natural for Black mayors, of whom all were Democrats except one in 2020, to defend their cities and embrace their racial identities in response to Trump’s posturing. This would require the rejection of respectability politics, but also the embrace of their Black urban identities in defiance of Trump—and racial distancing theory.
The Obamas in retirement are no longer the bellwether of Black politics. Instead, I suggest studying the behavior of Black mayors of America’s most populous cities to determine the present conditions under which a strategy rooted in identity politics versus racial distancing is more prevalent. To shed light on such conditions, I analyzed 232 political statements made by 31 Black mayors in response to high-profile police killings, BLM protests, and civil unrest. My analysis compares mayors’ rhetoric across two eras, with “Era 1” encapsulating responses to the officer-involved deaths of Michael Brown and Freddie Gray in 2014–2015, and “Era 2” encompassing reactions to the killing of George Floyd in 2020. I find strong evidence to suggest that, at least throughout the latter half of 2020, Black mayors routinely defied the politics of respectability in their public statements, speeches, and interviews.
Data and Methods
Characteristics of Cities and Mayors
I constrained my analysis to the 100 largest US cities using Ballotpedia, which uses 2013 population estimates to rank cities (Ballotpedia n.d.). Among these cities, all have populations greater than 200,000. Then, I consulted BlackDemographics.com, news media, and personal websites to determine which cities had a Black mayor during Era 1 or Era 2, plus each mayor’s political party affiliation, birth year, and gender. I identified 31 “vocal” mayors, i.e., those issuing at least one statement in response to the officer-involved killings and/or related BLM protests and civil unrest during an era (see Table 1). While five mayors held office during both eras, only Hancock of Denver was vocal in both time periods and was thus coded as “Hancock 1” and “Hancock 2.” Considering the limited overlap among vocal mayors across the two time periods, I do not attempt to identify change over time in individual mayors, but instead whether patterns in rhetoric are explained by the era in which statements were issued.
Table 1. Mayor and City Demographics, N = # of Statements
I then calculated the total number of years each city had experienced Black mayoral leadership since 1967 using information obtained from BlackDemographics.com, Ballotpedia, news media, and city government websites. I additionally consulted Census.gov to retrieve 2020 population estimates and racial composition data for each city. Finally, I obtained the vote share for Donald Trump in the 2016 and 2020 presidential elections for the county that each city resides in from Politico.com, which cites AP elections data.
Mayoral Statements Retrieval
The officer-involved deaths of Michael Brown on August 9, 2014, and Freddie Gray on April 12, 2015, inspired frequent demonstrations and civil unrest in Ferguson, Baltimore, and dozens of other cities during the fall of 2014 through the spring of 2015. Then, peaceful demonstrations for George Floyd, Black Lives Matter, and racial justice occurred in at least 1700 places throughout the summer of 2020. In some larger cities, civil unrest like clashes with police, rioting, vandalism, and looting also occurred (Haseman et al. Reference Haseman, Karina Zaiets, Procell, Petras and Sullivan2020). To locate mayoral statements about these incidents, which I sort into two eras, I used Nexis Uni, formally Lexis Nexis, which enables users to search a database of over 15,000 news, business, and legal sources including local, regional, national, and international newspapers, newswires, and transcripts of television and radio broadcasts.
I searched for the following, “Mayor Name” AND “Michael Brown” OR “Ferguson” OR “Freddie Gray” during “August 9, 2014–May 6, 2015” to locate relevant statements issued during “Era 1.” Then, to locate statements during “Era 2,” I searched Nexis Uni for “Mayor Name” and “George Floyd” during “May 25–December 31, 2020.” Searches involving mayors who gave numerous relevant statements that were covered widely by the news media returned a substantial number of documents, especially during Era 2. For instance, 1,809 documents were retrieved for Atlanta Mayor Keisha Lance Bottoms speaking about George Floyd. Every document description was meticulously evaluated to ascertain its relevance to the project. This was a time-consuming process, manageable mostly due to Bottoms being an outlier in the quantity of statements she issued and press coverage these received, along with mayors Lightfoot of Chicago and Carter of St. Paul. Most searches returned a few dozen to a few hundred results, which could be quickly scanned for relevance offsetting the time expenditure associated with vetting the results for the most prolific speakers. The entire text of all relevant documents was then downloaded and saved for coding, while superfluous documents were disregarded.
Additionally, in the few instances where a mayor served in office during both eras, it was common for searches to retrieve no relevant statements made in Era 1, but numerous statements issued during Era 2. Buffalo’s Byron Brown illustrates this: he apparently issued no public statements about the deaths of Michael Brown or Freddie Gray, while a search returned 167 results including 6 relevant statements about George Floyd from Era 2. Moreover, when Nexis Uni searches retrieved no relevant documents, I cross-referenced with Google searches including the same search parameters to make sure I was not overlooking items that might not have been reported in the news, like an official statement posted to a government website. This did not appear to be common and for the most part the Nexis Uni searches retrieved what I take to be a comprehensive, complete list of primary and secondary source material that included both full text of interview transcripts (rather than sound bites) and complete press releases, alongside news coverage of statements to the media. However, various news outlets may choose different key phrases from a mayor’s speech to quote in an article. To avoid variable media emphasis skewing the coding of statements, a sincere effort was made to identify, group, and combine/reconcile documents about the same public speech event so that the most complete statement was coded for each case of data.
Documents retrieved by these searches and included in my analysis therefore include all items with a direct quote attributed to a Black mayor in response to one of the specified incidents or related protests and/or civil unrest. These include newspaper articles quoting mayors, transcripts of speeches, press releases, interview transcripts from radio and TV, and the occasional tweet quoted in a news article. Statements that were retrieved during the search but superfluous to my research (about the city’s COVID-19 response, for instance) were categorized as neutral and omitted. I then categorized statements as interviews or speeches, with speeches including all scripted public statements, press conferences, news releases, and/or social media posts quoted by news media.
Urban Protest and Unrest
Urban centers draw substantial political activism surrounding officer-involved deaths, with the largest cities attracting the most civil unrest and/or violence (Astolfi 2020). A survey conducted by the Major Cities Chiefs Association (MCCA) identified 8,700 protests for racial justice occurring in 68 major cities and counties between May 25 and July 31, 2020 (Major Cities Chiefs Association 2020, 1). The MCCA report also suggests a relationship between city size, protest, and unrest, with Atlanta (pop. 532,695), Chicago (2.6m), and Dallas (1.3m), each reporting upward of 70% of protests involving civil disobedience compared to Buffalo (253,793), Newark (277,175), and New Orleans (347,772), which reported fewer than 10% of protests involving civil disobedience or violence (Major Cities Chiefs Association 2020, 6-8). Because information on the level of protest and unrest occurring across all cities in my analysis for both eras was unavailable, I include a dummy variable in my analysis that identifies cities with populations of 500,000 or larger, which serves as a proxy for the size of protest and amount of unrest cities experienced in response to the police-involved killings of Michael Brown, Freddie Gray, or George Floyd.
Coding Scheme: Embracing and Rejecting Respectability Politics
I articulate four key elements of each competing paradigm, which were identified, coded, and counted within each mayoral statement, down to the sentence level. Speech elements were coded as supporting respectability politics, i.e., as pro-respectability claims when they:
-
1) Emphasize proper individual behavior and morality. (Behavior)
-
2) Urge conformity rules and norms. (Conformity)
-
3) Use negative racialized stereotypes. (Stereotypes)
-
4) Encourage respect for rule of law and order. (Law)
Speech elements were coded as challenging respectability politics, i.e., as anti-respectability claims when they:
-
1) Affirm systemic racism. (Racism)
-
2) Support police reform and accountability. (Reform)
-
3) Affirm Black Lives Matter. (BLM)
-
4) Affirm civil right to protest and/or rationale for civil unrest. (Rights)
For instance, this is how a statement issued on May 30, 2020 by Melvin Carter, the mayor of Minneapolis’ twin city of Saint Paul, was coded:
That frustration, that pain, is real, and it’s legitimate (+1, affirms unrest). And to all of the people in our community who believe what I just said, who wholeheartedly need the world to hear that Mr. Floyd should be alive, that someone should be held accountable, and that we as a community, we as a culture, we as a society must do everything we can imagine to keep this from happening again, we stand with you (+1, supports accountability). I stand with you (+1, affirms BLM). Unfortunately, there are also those among us who would seek to use this moment, who would seek to use his death as an excuse, as a cover to agitate for the destruction of those same communities that have been most traumatized by George Floyd’s death (+1, encourages respect for the rule of law).
This means the “pro-respectability” code for this statement is 1. In the empirical analyses I control for the overall number of claims that are pro- or anti-respectability. Table 3 reports claim frequencies.
Table 2. Statement Claim Frequencies by Era
Table 3. Statement Type Frequencies and Mean Pro-Respectability Claims (M)
A research assistant, naïve to the hypotheses of the paper, was trained on the coding scheme and this person coded all collected statements for pro-respectability and anti-respectability claims. Then I (author), blind to the coding of the research assistant, independently re-coded a random sample of 20% of the statements (48/232) for use in estimates of the reliability of the research assistant’s statement coding (Lacy et al. Reference Lacy and Riffe1996). A two-way random effects model for estimates of consistency of the research assistant’s coding across statements revealed good reliability pro-respectability claims (ICC = .764), moderate reliability for anti-respectability claims (ICC= .518), and total claims (ICC=.682) in the statements (see Koo and Li 2015).
Data Analysis
I used linear mixed effects models to compare rhetoric across eras (Baayen et al. Reference Baayen, Davidson and Bates2008). Linear mixed effects models were used because the dataset violated several assumptions of more traditional general linear modeling (e.g., ANOVA, regression, t-tests). First, linear mixed models were used to reconcile the largely unbalanced nature of the dataset (see Tables 1-3), including differing numbers of mayors in each era, differing numbers of statements contributed by individual mayors, and differing statement lengths. Second, they were used to account for statements being nested within individual mayors, meaning some were not independent of one another and the associated error in modeling would be correlated within-mayor (Garson Reference Garson2013).
Linear mixed effects models were carried out in R (version 3.3.2) with the lme4 package (Bates et al. Reference Bates, Mächler, Bolker and Walker2014; version 1.1.12). To obtain fit estimates (β), I used the default REML (Restricted Maximum Likelihood) settings in lme 4. For model selection and testing, I used likelihood ratio tests (LRTs), with fitting with Maximum Likelihood (ML) fit settings. LRTs were carried out by contrasting a) comparing the target model to a reduced model that included all fixed and random effects except the fixed factor of interest to obtain a chi-squared (χ 2) test statistic (Barr et al. Reference Barr, Levy, Scheepers and Tily2013; Pinheiro and Bates Reference Pinheiro and Bates2000). Comparing a target model to a reduced (or null) model with all the same factors except the target variable using LRT is a common approach to variable selection in linear mixed model fitting to determine whether a given factor is important to the model and/or whether additional factors should be retained or not in a final model (Luke Reference Luke2017). Variables were considered significant to the model if the target model with the variable of interest was significantly different from a null model or reduced model (without that variable) at an alpha level of p < .05.
Results
Do political statements differ by Era?
To test whether rhetoric in mayoral statements changed over time, I used a linear mixed modeling approach (e.g., Barr et al. Reference Barr, Levy, Scheepers and Tily2013) with the number of pro-respectability claims in each individual statement as the dependent variable (i.e., Pro-Respectability Claims) and Era of the statement (i.e., Era 1 vs. Era 2) as the fixed effect independent variable of interest. The model additionally controlled for overall statement length (i.e., Statement Length = total number of pro- and anti-respectability claims in each statement) as a continuous fixed effects covariate, and variability in rhetoric of the individual mayors as random intercepts for each mayor (i.e., Mayor). Importantly, the random intercepts for each mayor account for the nested variance structure of the dataset, with differing numbers of coded statements attributed to individual mayors. This guards against the model being biased due to the imbalanced contribution of certain mayors, since they vary dramatically in the number of statements issued.
A likelihood ratio test comparing the hypothesized base Era model (Pro-Respectability Claims ∼ Era + Statement Length + (1|Mayor)) to a null model without the Era variable (Pro-Respectability Claims ∼ Statement Length + (1|Mayor)) showed a significant effect of Era (β= −.78, χ 2(1)= 5.93, p =.015). Specifically, Era 2 statements were still significantly lower in pro-respectability claims (M= 1.10) compared to Era 1 statements (M= 1.71), even after accounting for variance structure associated with the non-independence of statements of individual mayors and the influence of statement length on the overall model. Thus, it appears that Pro-Respectability rhetoric in Black mayors’ public discourse significantly decreased from Era 1 to Era 2.
Do mayor or city demographics further influence statements?
To explore whether mayor or city demographic factors explained additional variance in mayors’ rhetoric beyond that explained by Era (and therefore should also be retained in the model), I conducted several additional follow-up “contrast” models using a similar linear mixed model selection framework. Specifically, for each variable of interest, I compared a new, exploratory model with the added variable of interest as a fixed factor (e.g., Gender of mayor) to the base Era model (Pro-Respectability Claims ∼ Era + Statement Length + (1|Mayor)) and contrasted this exploratory model (e.g., model including Gender as a fixed factor: Pro-Respectability Claims ∼ Era + Statement Length + Gender + (1|Mayor)) with the main Era model (e.g., without Gender). I also tested for an interactive effect of the exploratory variable (Gender x Era).
First, I investigated whether the mayor’s Gender explained additional variance beyond that explained by Era. The model was not significantly improved by adding Gender (β= .15, χ2(1)= 0.34, p = .559) of the mayor or by adding an interaction term of Gender x Era (β= −.16, χ2(1)= 0.86, p = .769), suggesting that the gender of the mayor did not explain further differences in statement rhetoric across eras.
Next, I investigated whether the mayor’s Age (as continuous variable) was systematically related to rhetoric in the statements by Era. The model was also not significantly improved by adding this Age term (β= .008, χ 2(1)= 0.32, p = .567) or by adding an interaction term of city Age x Era (β=.05, χ 2(1)= 2.10, p = .147), suggesting that changes in Black mayoral rhetoric across era is not further explained by the age of the mayor making the statement.
I also investigated whether city size (< 500,000 or ≥500,000) was systematically related to rhetoric in the statements by Era. City size may be important as it was demonstrated to correlate with larger, more intense protests (Major Cities Chiefs Association 2020). However, the model was not significantly improved by adding this city Size term (β = .19, χ 2(1)= 0.48, p = .490) or by adding an interaction term of city Size x Era (β = .06, χ 2(1)= 0.02, p = .894), suggesting that mayors’ rhetoric was not further explained by whether their city is larger than 500,000.
I also investigated whether a mayor’s city having a majority Black population (>50%) was systematically related to their rhetoric in statements by Era. The model was not significantly improved by adding this city Race term (β = .35, χ 2(1)= 1.62, p = .203) or by adding an interaction term of city Race x Era (β = −.28, χ 2(1)= 0.24, p = .623). Alternatively, treating Black population as a continuous percentage rather than dichotomous majority (or not) did not change the results (Percentage Black: β = .006, χ 2(1)= 0.83, p = .362; Percentage Black x Era:β = −.003, χ 2(1)= 0.02, p = .889). Thus, the percentage of Black citizens in each city did not explain additional variance in Black mayoral rhetoric across eras.
I also considered whether the number of previous years the city had a Black mayor, a metric of incumbency, was systematically related to their rhetoric in statements by Era. The Era model was not significantly improved by adding this city Race term (β = −.009, χ 2(1)= 0.96, p = .327) or by adding an interaction term of city Race x Era (β = .006, χ 2(1)= 0.07, p = .786). Thus, Black mayoral incumbency of a given city did not explain additional variance in Black mayoral rhetoric across eras.
The final demographic I considered was percentage of vote for Trump in either 2016 (Era 1) or 2020 (Era 2), a metric of partisanship at the time. The model was not significantly improved by adding the percentage of Trump Vote (β = .02, χ 2(1)= 2.39, p = .122) or by adding an interaction term of Trump Vote x Era (β = −.006, χ 2(1)= 0.10, p = .756), suggesting that mayors’ rhetoric over this time was not further influenced by the local popularity of President Trump (i.e., partisanship).
Together, these results show that the base Era model is not significantly changed or improved by including additional mayor or city demographics, which leads to the conclusion that including them as further controls or explanatory factors in the model is unwarranted for this dataset.
Does the format of the statement further influence rhetoric?
In a final exploratory model, I considered the effect of the modality of statements, or whether a statement was given as a speech or an interview. The base Era model was significantly improved by adding this Format term (β = .70, χ 2(1)= 12.79, p = .0003), but not by adding an interaction term of Format x Era (β = .58, χ 2(1)= 1.41, p = .234). Thus, even after accounting for individual variation associated with each mayor and the effects of statement length, interviews contained fewer pro-respectability claims compared to statements given in speeches.
Discussion
Several observations follow from the empirical analyses of mayors’ statements in response to officer-involved killings, protest, and unrest. I find evidence of a significant shift in the rhetoric of Black mayors across time, with the number of “pro-respectability” claims per statement decreasing even after controlling for variance associated with the non-independence of individual mayors’ statements, as well as the length of statements. This analysis and the associated controls suggest that the rhetorical shift was characteristic of the population of mayors; it was not just driven by a few vocal mayors or by mayors being more vocal in one era compared to another. Moreover, mayors’ statements issued during interviews contained less pro-respectability rhetoric than those given during speeches, but since proportionally the number of speeches issued in Era 2 increased more than the number of interviews, this does not explain the shift in rhetoric over time.
Notably, exploratory model comparisons suggested that neither key demographic factors of mayors themselves (i.e., age/gender) nor major demographics of the cities in which they were elected (i.e., city size, partisanship, or racial composition) explained further variance in rhetoric regarding respectability or change in rhetoric across eras. On the one hand, this is surprising given that these are factors expected to explain political behavior and rhetoric (Abramowitz and McCoy Reference Abramowitz and McCoy2019; Gillespie 2009; Hajnal Reference Hajnal2007). On the other, even though my dataset contains the entire population of Black mayors in cities with populations over 200,000 and a large, representative sample of the relevant statements given, the number of Black mayors is (still) relatively small and the cities that are likely to elect Black mayors are still relatively homogenous. Given these limits it may not be surprising that these factors did not explain additional variance in my dataset above the factor of era. It is likely that these inherent limitations in Black mayor population size and homogeneity in cities electing Black mayors limited my ability to find statistically significant effects of specific demographics above and beyond Era. As the population of Black mayors and the diversity of cities that elect Black mayors increase, these factors will become increasingly important to study. In the meantime, what explains the observed shift in mayoral rhetoric?
One strong possibility is that mayors in 2020 were responding to the tone set by President Donald Trump, who repeatedly referred to protestors as “thugs,” and notoriously tweeted, “When the looting starts, the shooting starts,” (Wines Reference Wines2020). In response, Chicago’s Mayor Lightfoot spoke for roughly 40 minutes during a press conference in which she addressed civil unrest in Minneapolis, Chicago, and other major cities while strongly rebuking Trump’s rhetoric, which she referred to as “profoundly dangerous,” and, “the biggest dog whistle possible,” intended to “foment violence” (Bauer Reference Bauer2020). Then, before demanding an apology from the president, Mayor Lightfoot said, “I will code what I want to say [to Trump] and it starts with F and ends with U,” (Bauer Reference Bauer2020). Finally, when specifically asked by a reporter whether invoking profanity was befitting considering Michelle Obama’s dictum to “go high,” Lightfoot defied the politics of respectability, retorting, “Well, I’m not Michelle Obama,” (Bauer Reference Bauer2020).
This exchange between Lightfoot and Trump illustrates the profound political dilemma that “viral” police-involved killings present for Black mayors, as well as a potential strategy for navigating it. Nonetheless, controversial police uses of force and the unrest they incite have grown in symbolic importance as urban maladies. Such incidents accentuate the county’s intractable urban–rural divide by activating enduring negative stereotypes about the inhabitants and leadership of America’s cities.
President Trump has long sought to capitalize on anti-urban sentiment in his political speech, from his inaugural address in 2016 in which he referenced urban blight, i.e., “rusted-out factories scattered like tombstones,” to his 2020 Republican convention speech in which observed, “There is violence and danger in the streets of many Democrat-run cities throughout America” (Swanstrom Reference Swanstrom2017). Trump also pioneered assigning blame for police killings and urban unrest to mayors themselves, rather than broader structural problems. Prior to becoming commander-in-chief, Trump went on the offensive against Baltimore during the Freddie Gray protests by repeatedly tweeting that Baltimore was an example of, “corrupt government” and “a total mess,” despite billions in (wasted) federal investment (CBS Baltimore 2019).
By contrast, Obama’s condemnation of rioting in Ferguson and Baltimore in 2014 and 2015 was not framed as an indictment of urban life and leadership, but rather an admonishment of some bad apples’ behavior. Mayors in Era 1 could follow Obama’s tactic to embrace respectability politics without implicating their leadership as part of the problem. This makes sense considering the long shadow the “urban crisis” casts over American cities. Despite decades of social, political, and economic progress, cities are still routinely depicted by conservative politicians as crime-ridden hellscapes plagued by endless racial conflict and political corruption. It is perhaps coincidental that the first three “sanctuary cities” identified by Texas Governor Greg Abbott to receive busloads of migrants from his Southern border—New York, Washington D.C., and Chicago—were all led by Black Democrats. It is also perhaps not: the observation that many of America’s largest cities with complex problems are led by Black mayors is not lost on their detractors. Nelson and Meranto warned of such “dilemmas of power,” since the precise characteristics of cities that create opportunities for Black leadership also create barriers to effective governance and urban reform, namely white urban exodus, a declining tax base, and an inner-city constituency with fewer resources and more needs (Reference Nelson and Meranto1977, 336). The election of a Black mayor, rather than signaling racial progress, may therefore be viewed as, “the onset of black takeover of bankrupt cities consumed by social conflict, physical decay, and enormous financial problems” (Nelson and Meranto Reference Nelson and Meranto1977, 337).
However, the trajectory of Black mayoral leadership in major American cities has defied Nelson and Meranto’s cautionary tale. Despite persistent challenges, urban life has improved on key indicators during the 21st century. The homicide rate, symbolic and symptomatic of urban dysfunction in the 1980s and 1990s, plummeted sharply and broadly during the early 2000s, including in cities with sizeable Black populations and decades of unbroken Black leadership like Atlanta, Birmingham, and Washington, D.C. Cities are also viewed as ascendent in global policymaking and leadership in environmental, technological, and social innovation (Swanstrom Reference Swanstrom2017; Bramwell Reference Bramwell2020; Kaufman and Sidney Reference Kaufmann and Sidney2020). Additionally, Black mayors have made inroads as leaders of majority and plurality-white cities like San Francisco, Denver, and Dallas; in 2020 fully one-fifth of the nation’s most populous cities were led by Black mayors, representing 19.5 million Americans! Black mayors are poised to emerge as vocal representatives—indeed champions—of race and place in the current era.
However, as I write this after the second anniversary of Floyd’s death, there are reasons to expect respectability politics to endure, especially on certain issues. As Bunyasi and Smith (Reference Bunyasi and Watts Smith2019) have argued, “even in the BLM era” Black Americans embrace the logic of respectability politics while seeking to distance themselves from (extra) marginalized Black subgroups like women, LGBTQ, and the formerly incarcerated. Then, support for Black Lives Matter has “plummeted” since the summer of 2020, especially among Black Americans (Rahman Reference Rahman2022). According to a study by the University of Massachusetts Amherst, support for Black Lives Matter’s goals declined sharply from 48% in April 2021 to 31% in May 2022. Among Black respondents in particular, support for Black Lives Matter fell from 67 to 56 percent (Rahman Reference Rahman2022).
Related to this, mayors like Lori Lightfoot of Chicago, Eric Adams of New York City, and London Breed of San Francisco have frequently embraced tough rhetoric and policies in the years since Floyd’s death in response to problems associated with homelessness, illicit drug use, and violent crime in their cities. For instance, New York City’s post-pandemic crime spike is speculated to have bolstered Eric Adams, a former police officer who embraced public safety elements of respectability politics during his successful 2021 bid for mayor (Associated Press 2021). Then, London Breed notoriously characterized rampant crime, homelessness, and drug use in the Tenderloin as “the bulls---” destroying San Francisco during a press conference in December 2021, where she then called for an aggressive law enforcement response (Johnson and Schneider Reference Johnson and Schneider2021). Finally, fierce Trump opponent Lori Lightfoot, once lauded as a leading police reform advocate, has shifted much of her administration’s focus since 2020 on surging gun violence and car-jackings in the Chicago, promising tougher penalties for crimes involving gangs, guns, and/or violence (Bradley Reference Bradley2021).
Moreover, Black citizens are no longer the majority constituency in Atlanta and D.C., or the largest racial group in Philadelphia, and sizeable Black populations in cities like Chicago, New York, Oakland, and Los Angeles have shrunk in recent decades (Blow Reference Blow2021). At the same time, the Black population has increased in cities like Houston and Dallas, enhancing opportunities for Black leadership to emerge in novel places. In the post-Obama era, majority-Black cities like Birmingham and Baltimore continue to elect Black mayors, while majority-white cities with Black strongholds like New York, Chicago, and Dallas have demonstrated a willingness to elect Black mayors. Additionally, will more Black Republicans seek election in conservative places, following Sylvester Turner in Houston? Black mayors will likely continue to increase their power and visibility if they are successful in building racial coalitions. To this end, racial distancing may be a winning position on some issues, like public safety and homelessness (Campanile and Beeferman Reference Campanile and Beeferman2021). I imagine that political expedience will inform Black mayors’ decisions, as will the political behavior and attitudes of presently sorted white urban democrats, and less well-sorted Latinos and Asian-Americans.
Finally, as the field of Black urban leaders increases and diversifies, what type of racial strategies might they adopt based on personal and city demographics? Although relevant demographics did not statistically explain additional variance above and beyond era in this dataset, statement format did. A broader sample of mayoral statements, including additional sources like social media posts collected over a longer time span, and/or on an extended number of subjects, could be even more informative as to how mayor and city demographics further influence their rhetoric. Given this, future research should examine the diversity of perspectives of Black mayors considering their distinctive backgrounds, the unique racial politics and policy challenges confronting their cities, and the sitting president’s rhetoric and record on racial issues.
Acknowledgments
I would like to thank Brighton Robbins for her help gathering, organizing, and coding statements and Dr. Daniel C. Hyde for consulting on the statistical analyses.
