Ethics of Care Leadership, Racial Inclusion, and Economic Health in the Cities: Is There a Female Leadership Advantage?

Social inclusion, respect, care, and justice are at the heart of the ethical city.

(Barrett et al., 2016, p. 5).

Abstract

Growing evidence suggests the presence of a female leadership advantage (FLA), such that women leaders tend to be associated with more effective outcomes in uncertain conditions. However, mechanisms linking women's leadership to effective outcomes are less well understood. We integrate FLA insights with ethics of care philosophical framework to conceptualize how women leaders achieve effective outcomes in the context of the urban revitalization crisis in the United States. We propose and empirically test the mediating role of ethics of care leadership in the relationship between women mayors and economic health of their cities. We used data from the Urban Institute that includes 272 United States cities and measures of variables in our conceptual model at five points in time spanning 36 years (n = 1185 city-year observations). We capture ethics of care leadership focused on racial inclusion with an index measure of a city’s racial spatial segregation, homeownership gap, poverty gap, and education gap, and we capture economic health with an index measure of a city’s employment growth, unemployment rate, housing vacancy rate, and median family income. We found that female-led cities were associated with better economic health, and this association was mediated by female-led cities’ association with greater racial inclusion. Ethics of care leadership appears to be one pathway through which a FLA manifests itself in the context of the urban revitalization crisis. This underscores the importance of city leadership that balances social and economic prerogatives. Implications are discussed.

Economic health of U.S. cities varies both between and within them. When city leaders manage economic development, they typically have at their disposal established research to guide their efforts (Kodrzycki & Muñoz, 2010; Wolman et al., 2004). Largely absent from that literature are strategies for boosting inclusion of marginalized groups as a part of urban renewal. This is needed because as cities expand their economies, wealth inequalities within them tend to expand too (Kochhar & Fry, 2014). When benefits from economic growth are unequally distributed across residents, excluded people of color bear the social consequences (Chetty et al., 2018; DiTomaso et al., 2007).

Social exclusion … involves the lack or denial of resources, rights, goods and services, and the inability to participate in the normal relationships and activities, available to the majority of people in society, whether in economic, social, cultural, or political arenas. It affects both the quality of life of excluded individuals and the equity and cohesion of society as a whole. (Levitas et al., 2007, p. 9)

Exclusion from opportunities has been stratified by race in U.S. cities (Rothstein, 2017). Drawing from multiple theory lenses, we propose a model in which leader gender is associated with economic health of cities, and this relationship is mediated by ethics of care leadership focused on racial inclusion (see Fig. 1). Studies have mostly surveyed if economic growth leads to inclusion (Einstein et al., 2017; Shearer et al., 2017), and the results are mixed. Some cities have strong economies but low inclusion, while others show high inclusion and poor economic health, and yet in other cities, there is no connection (Benner & Pastor, 2012; DeSouza-Briggs et al., 2015). In the present research, we examine if racial inclusion is related to economic health. In this way, more residents contribute to the local economy, but city leadership is needed to promote inclusion in the first place.

Fig. 1

Conceptual model of a female leadership advantage as a pathway to an ethical city

Because our research context is the revitalization of urban environments, the leaders we examine are mayors. Leadership, a process of motivating effective group behavior (Haslam et al., 2020), is vital in amorphous social contexts in which practices from predictable times are unlikely to apply (Hannah et al., 2009). Leadership research has offered few answers to the urban crisis (Florida, 2017) and the role mayors might play in it. Literature on a female leadership advantage (FLA) suggests that women, in comparison to men, might be more effective leaders when organizations are in crisis (Helgesen, 1990; Post, 2015; Post et al., 2019; Rosener, 1990). Drawing from this literature, we examine if and how women-led cities might be more effective than men-led cities in terms of economic health.

We propose that this relationship is mediated by ethics of care leadership, manifested by greater racial inclusion. In line with our focus on cities, we take a sociological perspective on racial inclusion and consider it at the group level (Tilly, 1998). Thus, inclusion represents unimpeded access to economic opportunities and other benefits for city residents regardless of their race or ethnicity. Because a key exclusion fault line in the U.S. falls along race and ethnicity (Guynn & Schrotenboer, 2021; Roberson, 2019), we examine inclusion of people of color, represented as anyone who self-identified in the U.S. Census as non-Hispanic White, i.e., Black or African American, Asian, American Indian and Alaska Native, Native Hawaiian and other Pacific Islander, other race, two or more races, or Hispanic or Latino.

Ethics of care was developed in response to traditional moral theories ignoring feminist critiques of them (Baier, 1994; Walker, 1989). Caring is critical to survival, as infants and children fully depend on others to care for them (Kittay, 1995), to say nothing of caring being instrumental for people to live in social harmony. That is, “caring can be viewed as an activity that includes everything that we do to maintain, continue, and repair our world so that we can live in it as well as possible” (Tronto, 2013, p. 19). Women tend to engage in caring more so than men (Barnes, 2012), and it has been established that women consider care a moral virtue (Hanen & Nielsen, 1987; Kittay & Meyers, 1987). Yet, for example, Pincoffs (1986) lists 221 personal virtues, but care is missing. Similarly, MacIntyre (1999) recognizes the importance of human dependency on one another yet fails to acknowledge care as a moral virtue. Feminist philosophers argued for the consideration of care as an ontological virtue and pushed for its inclusion in the epistemology of moral theories, known as ethics of care today (Held, 2006, 2015). Ethics of care framework enables women to better discern their voices and contributions in creating a caring society across domains of activity (Glenn, 2000; Tronto, 2013).

Bridging this theory and leadership research, we introduce ethics of care leadership (ECL), a framework in which ethics means that care is wanted and unobtrusive (Glenn, 2012), care is the value (Held, 2015), and leadership is the practice of attentiveness toward others, responsiveness to their needs, and cultivating caring relationships (Robinson, 2011). We propose that women mayors, compared to men, are more likely to achieve economic health by fostering racial inclusion in their cities because they are more likely to embrace ethics of care leadership.

We pooled data from multiple public sources into a unique dataset that includes 1185 city-year observations, spanning 272 U.S. cities and 971 unique mayors, over 36 years with five measurement points (1980, 1990, 2000, 2013, 2016). Results from complementary analyses supported our hypotheses. Compared to cities with male mayors, those with female mayors were associated with healthier economies, and this finding was mediated by greater racial inclusion.

This study makes the following contributions. First, scholars have made calls for greater involvement of leadership research in addressing social issues (Walsh et al., 2003). We answer this call by bridging FLA research and ethics of care moral philosophy and by studying it within the general context of the urban revitalization crisis in the U.S. In this context, a crisis is defined as a co-occurrence of high-impact issues that can create ripple effects if left unaddressed (Hannah et al., 2009; Weick & Sutcliffe, 2001). FLA studies on leadership in crises have mostly focused on organizational crises (e.g., product recalls, financial downturns) (Post et al., 2019), where prompt action is needed (e.g., instituting emergency measures after a market crash) (Kahn et al., 2013). Though valuable, findings from those crisis contexts might not translate to effective leadership qualities during an ongoing societal crisis. Unlike organizational crises, the constellation of factors contributing to a chronic societal crisis form over time to create enduring tensions for which an immediate fix is not readily discernible (Weick, 1988). Exclusion by race in U.S. cities over decades has led to segregated experiences across the nation (Rothstein, 2017), giving rise to an urban crisis characterized by racial gaps in spatial segregation, homeownership, poverty, and education (Krysan & Crowder, 2017). We leverage this ongoing social crisis to answer a call made by Post and colleagues (2019) for more research on gender and leadership effectiveness in various crises contexts.

Second, one problem in the FLA literature is that studies on gender and leadership effectiveness tend to lack explanatory mechanisms linking gender and outcomes (Hoobler et al., 2018; Post & Bryon, 2015). As reviewed by Hoobler et al. (2018), leaving pathways unspecified that may elucidate ways in which women leaders achieve outcomes has led critics to label FLA research a “body count” (p. 2484) or “subjective tokenism” (p. 2490). By conceptualizing and testing a mediation model in which a FLA unfolds through ECL focused on racial inclusion, we contribute to a still relatively open question of what broad gender differences might inform women’s leadership effectiveness.

Third, the design of this research has reasonable ecological validity because it is based on real-life data from the urban crisis in the U.S. as it unfolded over nearly four decades. Multiple and objective measures of the criteria add to construct validity and robustness of the findings.

Literature Review and Theory Development

A Female Leadership Advantage

Gender differences in leadership styles and their effectiveness were first realized following two ground-breaking studies in 1990 (Helgesen, 1990; Rosener, 1990). Both foresaw an opportune juxtaposition of two trends, dubbed a “fortunate confluence” Helgeson (1990, p. 14): business upheaval of the 1980s and the growing influx of women into management roles. The former called for exploration of new leadership styles because the standard practices from more predictable times were unlikely to apply and the latter movement provided fruitful answers. As explained by Helgesen (1990), “As historic outsiders to such positions, women often had fresh eyes to see what was no longer working and to identify new solutions” (p. 14). Similarly, Rosener (1990) predicted that “By valuing a diversity of leadership styles, organizations will find the strength and flexibility to survive in a highly competitive, increasingly diverse economic environment” (p. 125). Instead of enumerating alleged handicaps of women in business, as was common practice, these studies focused on what women could offer in the changing workplace.

Compared to an earlier study by Mintzberg (1973), in which five male executives reported being essentially “company men,” Helgesen (1990) and Rosener (1990) found that women approached work differently. Mintzberg (1973) had documented that men tended to work long hours (and suffered accordant family/social drawbacks), rely on a top-down chain of command, monopolize information at the top, have a deep but narrow focus, develop skills to follow procedures, value conformity to corporate norms, and focus on goals, competing, and wining. In contrast, Helgesen (1990) and Rosener (1990) found that women tended to share information, invite participation in decisions making, engage in relationship building, ensure others felt included, and take a long-term focus. Female leaders did not perceive themselves as “company women” following typecast norms (see also Kanter, 1977). Instead, they saw themselves and their colleagues as humans with varied backgrounds who react differently to work challenges. In both studies, women achieved business results, fulfilling the role of agents of transformation. Helgeson (1990) coined the phrase female leadership advantage (FLA) to describe her findings, and Rosener (1990) summarized her findings on gendered leadership characteristics as tending to be transactional for men and transformational for women (per Bass, 1985).

Although methods (e.g., small sample sizes) in these early studies were criticized (Vecchio, 2002), the idea of gender-driven leadership effectiveness gave rise to many studies in the following decade. Generally, this research found similar results, i.e., men tended to espouse power, make top-down decisions, guard information, have a transactional view of employees, exhibit competitiveness, assertiveness, and a willingness to take risks, whereas women leaders seemed to embrace inclusiveness, information sharing, caring, kindness, and compromise (Carli, 1990; Carr-Rufino, 1993; Jaffee & Hyde, 2000; Lenney, 1991; Rosener, 1995; Schein et al., 1996; Yukl, 2002). Eagly and colleagues meta-analyzed much of this research, and results largely affirmed a FLA (Eagly & Johnson, 1990; Eagly et al., 1992, 1995, 2003). While these findings do not imply that all women espouse caring values and men do not, they do indicate the existence of gender-based differences that can shed light on women's leadership approaches.

Critics argued that evidence behind a FLA is insufficient to be convincing (Adams, 2016; Pletzer et al., 2015; Vecchio, 2002), and debate ensued (Eagly, 2007, 2016; Eagly & Carli, 2003a, 2003b; Tavris, 2001; Vecchio, 2003). Social role theory (Eagly, 1987; Eagly et al., 2000) suggested that a FLA can turn into a disadvantage when sex roles are stereotyped and enacted against the social stereotype (Eagly, 2007). Glass-cliff research (Ryan et al., 2016), on the other hand, suggested that women, more so than men, are appointed to leadership during times of uncertainty simply because there is less competition for precarious positions (Bruckmüller & Branscombe, 2010). Together, being a woman leader has been viewed as an advantage, a disadvantage, or neutral (Post, 2015; Rosette & Tost, 2010).

We suggest that a FLA can be at least partly ascribed to gender differences in moral values. In mainstream moral psychology, values such as independence, self-determination, and individual rights are prioritized, and there appears to be “a systematic devaluing of notions of interdependence, relatedness, and positive involvement in the lives of others” (Robinson, 1999, p. 7). Though the former values are important, Walker (1989) argues that they represent “the moral agent” showing “how” the rest of us ought to act (p. 21, emphasis original). These values espoused by dominant figures in a social order (e.g., judge, bureaucrat) largely elude the daily experiences of women (see Barnes, 2012), such as caring for family, building and maintaining relationships, and addressing needs of those who rely on community members for aid. Evidence from multiple sources, including psychoanalytic practice, has converged in support of care as a moral value, and studies suggest care tends to be more valued by women than by men (Held, 2006, 2015; Jordan, 1997; Jordan et al., 1991; Miller, 1976).¹ Put simply, applied psychology research has virtually never found that women exhibit more interest, on average, than men in pursuing economic gain at the expense of community, nor that men are more inclined, on average, than women to focus on care at the expense of business interests ( Gilligan, 1982; Taylor et al., 2000; Wood & Eagly, 2002). These gender differences in the value placed on care as a moral virtue seem likely to lead to differences in the degree to which ethics of care leadership is espoused.

Alternatively, strengthening stereotypes associating women as communal (Eagly et al., 2020) may prompt women to intentionally conform to norms and avoid backlash by leading in more caring ways. Further, selection-based views of leadership might suggest that women who support, express, or embrace ethics of care will be more likely to be promoted to leadership roles because this style conforms to social role stereotypes (Eagly & Koenig, 2021; Vial et al., 2016). Whether differences in ethics of care leadership arise from gender differences in moral virtues, as we suggest, or from a blend of reasons suggested above (e.g., social norms, selection), we suggest that women leaders are more likely to embrace ethics of care leadership (Elleyl-Brown & Pringle, 2018; Pullen & Vachhani, 2021).

Ethics of Care Leadership and Racial Inclusion

Ethics of care leadership (ECL) is an activity that involves expenditure of physical and mental energy; it is not a mere attitude, such as “I do not care for this or that,” nor is it just a warm feeling for someone. ECL is a behavior aimed toward the well-being of a relationship and those in the relationship (Ruddick, 1998). To clarify, this does not include caring for all of humanity all the time (Held, 2010; Okin, 1998). Instead, it is caring for those in a particular personal or social relationship, which can comprise of any number of people, including an entire community. For this reason, the interests of those caring and those who are cared-for are interwoven rather than independent or competing, and caring is voluntary rather than contractual (Glenn, 2012; Robinson, 2011a). This literature rejects the view that caring is simply a biological necessity as a means of survival (see Tronto, 2007; Collins, 2015) and, hence, not a moral virtue. In contrast, acts of caring can be transformative, as they help people (re)gain capacity to live in mutually beneficial relations with others (Petterson, 2008). For example, when in crisis, people can feel overwhelmed and uncertain of how to move forward. Caring leadership is what many in the community look toward for a viable path out, as well as seeking help with the discomfort.

Racial inclusion is aimed at building a more equitable society, and those efforts can benefit from ECL. Generally, there are three aspects of inclusion (Bulger et al., 2021). First, inclusion is more than an outcome to strive for; it is a process of connecting people and opportunities, especially excluded communities (Bates & Seddon, 2008; Silver, 2015). Second, relatedly, inclusion scholars and practitioners have argued for a holistic approach to studying it versus focusing on single issues in isolation (Krishna & Kummitha, 2017). Finally, “to bring inclusion to fruition, there needs to be a purposeful vehicle of initiative through which a process of social inclusion can be operationalized” (Bulger et al., 2021, p. 2). ECL, with its emphasis on caring, might be that vehicle of change toward greater inclusion in urban areas.

Ethics of care has been advocated for in the political arena (Tronto, 2010), which transitioned into an emerging literature on ethical cities (see Barret et al., 2016). In an ethical city, leadership strives toward inclusive community life, both economically and socially. This goal is in relative contrast to prior conceptions of city leadership. For instance, in a neoliberal city, a citizen is seen as the consumer and city governance is modeled after a for-profit enterprise (Hackworth, 2006). Hence, typical mayoral priorities focus on economic development (e.g., securing investments, boosting new ventures, supporting real-estate developments), maintenance and development of infrastructure (e.g., roads, bridges, water/sewer systems), offering reliable public services (e.g., mass transportation, school systems, policing, prisons), and, if possible, boosting consumer spending (see NLC, 2022). These traditional mayoral activities have helped cities amass substantial resources. However, those resources are often found in segregated allotments, signaling exclusion of people of color from opportunities (Oxfam, 2017). Business growth is critical as a tide that raises all boats, but markets can be exclusionary of some communities in this uplifting process (Poethig et al., 2018).

We highlight a new perspective on mayoral priorities associated with ECL that differ from positions focusing exclusively on revitalizing cities via an influx of business development (Weinberg, 1996). We propose that mayors who embody ECL go beyond revenue-generating and budgeting by also recognizing and accommodating the diverse interests and groups embedded in urban life. Under ECL, governance is mutually reinforcing as mayors make intentional choices to rebuild communities and make their cities more inclusive (Barrett et al., 2016). Deliberate effort toward inclusion is needed because sole reliance on markets to fix the problem has been ineffective: “Urban inequality remains an endemic and worsening experience for many. Indeed, … inequality is the key challenge facing our cities” (Barrett et al., 2016, p. 3).

We extrapolate research that finds that women value care as a moral virtue more so than men to suggest that women leaders are more likely to embrace ECL. As a result, compared to men, women mayors are more likely to foster greater racial inclusion in their cities through the following set of practices. First, they are more likely than men to generate greater and more fruitful engagement with those they lead (Belenky et al., 1997; Bond et al., 2008; Hrdy, 2009; Post, 2015; Post et al., 2022). ECL follows rules as an important baseline, but cultivating relationships moves the community forward together (Ranieri & Ramos, 2013). On average, women tend to be more engaged than men in their communities, centering their involvement around volunteerism and activism (Bratton & Haynie, 1999; Rosenthal, 1998a). For instance, research indicates that women leaders actively oppose toxic waste storage near residential areas (Brown & Ferguson, 1995), work with vulnerable city youth (Abrahams, 1996), and advocate for affordable housing (Saegert, 1989). For urban revival to be effective, buy-in from fragmented community members is crucial. Because inclusion cannot be forced, it is unclear how men’s tendency to use hierarchy and power to compel compliance would stimulate inclusion across racial lines. Relatedly, in approaching dilemmas, men tend to be more concerned with justice—interference with others’ rights in impermissible ways—than with care (Haidt, 2012; Rawls, 1999). Thus, men tend to make “decisions using rules, regulations, and traditional ways of doing business” (Bart & McQueen, 2013, p. 97). This is important for maintaining order, but it does little to nurture racial inclusion.

Research in political science, public policy, and gender studies corroborates these links. For instance, men and women legislators report approaching their roles differently (Rosenthal, 2002). Men tend to approach equity issues with an agenda-improvement mindset by supporting incremental changes to existing programs, fixing issues within extant bureaucratic procedures (Boles, 1991), and allocating funds to regulatory bodies (Kathlene, 1989; Kathlene et al., 1991). In contrast, women politicians are more inclined toward agenda-building, such as “reshaping the agenda” (Dodson & Carroll, 1991, p. 1), “transforming the agenda” (Swers, 2002, p. 260), offering new views and ways of doing things (Gerrity et al., 2007; Lunneborg, 1990), and prioritizing novel inclusive policies (Weikart et al., 2007). These differences have been documented in state legislatures (Poggione, 2004; Rosenthal, 1998b), Congress (Rosenthal, 2002), and in comparisons of men and women city managers (Fox & Schuhmann, 1999, 2000; Funk & Molina, 2021).

Another factor through which women mayors might foster greater inclusion is by being more attentive to human differences. Compared to men, women leaders tend to be more “sensitive, attuned, and responsive to moments of differences, and feel responsible for working with the differences” (Cunliffe & Eriksen, 2011, p. 1438). This suggests that women leaders are more likely than men to consider multiple stakeholder perspectives, learn from multiple constituencies, and course-correct toward inclusion and community building (Pratto et al., 1997; Rinehart, 1991; Rosener, 1995). This type of broad collaboration enables women leaders to leverage knowledge and skills from multiple stakeholders and make decisions by drawing from a richer source of information; in this way, women are often seen as relational leaders who value interdependence, benevolence, and tolerance (Adams & Funk, 2012; Clement, 2018). Together, women leaders form a “we are in this together” platform underlined by a “because we care” value (Borgerson, 2018; Haidt, 2012; Haslam et al., 2020). By nurturing collaboration across city constituencies, women leaders galvanize effort toward racial inclusion.

Organizational leadership research on CEOs and board members likewise supports these connections (Loyd et al., 2013; Peterson & Philpot, 2007). For example, corporate women leaders, compared to men, are more likely to prioritize broad initiatives related to inclusion, equity, and social responsibility (Bilimoria & Wheeler, 2000; Dezso & Ross, 2012; Glass & Cook, 2017; Hafsi & Turgut, 2013; Seto-Pamies, 2013), and organizations with female top executives have fewer discrimination lawsuits (Abebe & Dadanlar, 2021; Dadanlar & Abebe, 2020). Similarly, studies on politics, policy, and gender suggest that women legislators, in comparison to men, devote more effort to wide-ranging social programs that provide aid to underprivileged communities (Pratto et al., 1997; Swers, 2001) and co-sponsor policies that seek to more evenly spread wealth (Poole & Zeigler, 1985; Shapiro & Mahajan, 1986; Swers, 2005).

Racial inclusion in cities manifests in racial spatial segregation, as well as racial gaps across homeownership, poverty, and education (Bratton & Haynie, 1999). To illustrate how mayors can improve inclusion, we identified several real-life examples of policy making. Levitt et al. (2018) outlined reporting standards that called for “quotes or excerpts to augment data” (p. 37) and to “demonstrate that findings are grounded in evidence (e.g., using quotes)” (p. 36).

In addressing inclusion, not all mayors pursue the same action, and some choose inaction. For example, of the 1621 mayors and officials who perform mayoral functions of U.S. cities, towns, and civil divisions with populations over 30,000 in 2021 (CAWP, 2021), only 143 signed the “Mayors’ Contract on Racial Equity,” publicly committing to promoting racial equity, compassion, and justice for all city residents.² Although women constituted only 25% of U.S. mayors in 2021 (CAWP, 2021), they represented 40% of the mayors who signed this document. This suggests that women mayors, in comparison to men, appear more likely to prioritize racial inclusion in their cities.

Miami Mayor Francis Suarez refused to sign the inclusion pact. Pointing to differences in approaches to inclusion, he stated: “We would love for everyone to be prosperous … [but] we’re trying to find the right balance … fostering innovation to empower people to be able to do that on their own.” In contrast to Mayor Suarez’s approach, Mayor Libby Schaaf of Oakland, CA announced a privately funded project that afforded low-income families a chance to secure steady income for 18 months—no strings attached (BondGraham, 2021). When explaining her policy choice, she stated “We believe that poverty is not personal failure, it is policy failure.”

Mayor of Chicago, Lori Lightfoot, introduced a five-year plan to tackle racial spatial segregation by increasing education and outreach for voucher holders to encourage minority families to live in higher opportunity areas (Collinson & Ganong, 2018). She further introduced incentives and re-zoning initiatives to boost affordable housing by encouraging multifamily unit developments near transit hubs (Lightfoot, 2021). Mayor Muriel Bowser of District of Columbia similarly sought to promote inclusion in her city by creating a task force to identify actionable steps toward greater minority homeownership. She explained that goal was to “…address decades of racially discriminatory policies and practices that have hampered access to one of the most significant ways to build wealth for Black residents. … [I] intend to chart a path to rectify these problems … and to set a 2030 goal for Black homeownership.”³

Taking another approach, Mayor Amy Shuler Goodwin of Charleston, WV partnered with a charity to secure funds for a city plan aimed at improving lives of those residing and working on the west side of Charleston by increasing access to housing and jobs (NLC, 2022). Mayor Regina Romero of Tucson, AZ allocated federal funds for youth employment and low-income housing, explaining “What we’re trying to do with American Rescue dollars is bring about transformative change through investments in the causes of poverty” (Murakami, 2022). St. Louis Mayor Tishaura Jones allocated those same funds for cash payments to residents of color, explaining “It’s a priority for me to address the historic disinvestment of Black and Brown communities…When we provide support to those who need it most, we will stabilize our communities.” (Murakami, 2022). Taken together, we offer the following three hypotheses:

H1::

Cities led by female mayors are likely to be associated with better economic health compared to those led by male mayors.

H2::

Cities led by female mayors are likely to be associated with greater racial inclusion compared to those led by male mayors.

H3::

Racial inclusion of the city mediates the relationship between mayor gender and economic health of the city.

Method

Data Sources

We used the “Inclusive Recovery Report” publicly available data released by the Urban Institute (Poethig et al., 2018).⁴ The Urban Institute collected data on U.S. cities with a population of 100,000 dating back to 1980. Only incorporated cities with a municipal government were included in their dataset because these cities have leaders (i.e., mayors) with authority over economic and social policies. Measures of inclusion and economic health were available for 274 cities at five points in time (1980, 1990, 2000, 2013, and 2016).

This resulted in a dataset of 1370 city-year observations. Of these, we identified the mayor for 1185 observations (86.5%). Of the missing observations, 173 (93.5%) related to 1980, 1990, or 2000, as online archival capabilities were lacking compared with 2013 and 2016.⁵ We analyzed the entire dataset on both criteria (racial inclusion; economic health) using marginplot from VIM 6.1.1.tgz package in R (version 4.2.1) and confirmed that missing data were randomly distributed (see Fig. 2). Hence, the coefficients and standard errors are unlikely to be biased.

Fig. 2

Missing data diagnostics. Measures of inclusion and economic health were available from the Urban Institute for 274 cities at five points in time (1980, 1990, 2000, 2013, and 2016). This resulted in 1370 city-year observations. Of these, we were able to identify the mayor for 1185 observations (86.5%). We analyzed entire dataset on both dependent measures (racial inclusion index, left panel; economic health index, right panel) using margplot from the VIM 6.1.1.tgz package in R (version 4.2.1). In the left panel, the green data points represent observations of racial inclusion index separated by whether the mayor was identifiable (left, identified; right, unidentified). In the right panel, the blue data points represent observations of economic health index separated by whether the mayor was identifiable. The data across both DV’s in the right column (unidentified mayor) appear to be randomly distributed, confirming the data are missing completely at random

Measures

Mayor Gender

Research assistants blind to hypotheses coded the mayor’s gender, a predictor in our model, as female or male from biographies and pictures available online (e.g., city website, articles, obituaries, LinkedIn). Of the 971 mayors in the dataset, 172 were women (17.71%).

Racial Inclusion

A mediator in our model, this construct represents the ability of residents of color to contribute to community life and share in its benefits. The Urban Institute defines residents of color as those who do not identify as non-Hispanic whites, according to their self-identification in the U.S. Census.⁶ For each indicator variable included in the racial inclusion construct, raw scores were collected and converted by Urban Institute to z-scores, reflecting the relative score of each city on the indicator variable in comparison to the full 274 city sample, with 0 as the mean and 1 as the standard deviation. Racial inclusion indicator variables were chosen by the Urban Institute to “reflect policy areas over which city leaders have some control (e.g., housing, job quality, education) so that the indices can directly inform local policy change,” and each indicator was weighted equally by Urban Institute. Thus, the racial inclusion measure is a comparative index, or average Z-score, across four indicator variables, and it represents the relative racial inclusion score compared to other cities included in the sample.⁷ The z-score was computed by Urban Institute such that a high score represents greater racial inclusion.

Spatial Segregation

The first indicator variable of racial inclusion is spatial (residential) segregation by race (Charles, 2003) measured by a dissimilarity index (Benner & Pastor, 2015; Logan & Stults, 2011).⁸ This index reflects the relative distributions of two groups (people of color and whites) across neighborhoods in the same city. It ranges between 0 and 100, indicating the percentage of one group needing to move across neighborhoods to be distributed equally to the second group, where 0 indicates both groups are distributed proportionally across all neighborhoods, as this index was used in prior research (Parisi et al., 2015).

Residential segregation represents spatial circumstances that curtail opportunities for participation in civil society for those living in them (Saatcioglu & Ozanne, 2013). It is a limiting context because “a person’s patterns and norms of behavior tend to be shaped by those with which he or she has the most frequent or sustained contact and interaction” (Wilson, 1987, p. 61). The result can be a clustering of neighborhood attitudes and practices that deepen exclusion of local residents from the rest of the community (Bolt et al., 1998). Whether reasons for spatial exclusion are misguided housing policies (John, 2007), market forces in the form of prejudicial real-estate practices (Rothstein, 2017), and/or personal biases (Rabinowitz & Perry, 2002), residential segregation comes with a shortage of opportunity, has high barriers to exit, and it persists, compounding inequalities (Logan & Stults, 2011; Lyons & Pettit, 2011). Mayors can take steps to mitigate residential segregation with a mix of policy initiatives, such as foster housing density near mass transit, employment, and services, sharing information about voucher programs for education if they exist, and encouraging mixed-income communities (see also Florida, 2017; Rothstein, 2017; Taylor, 2019).

Homeownership Gap

The second indicator variable of racial inclusion is a racial homeownership gap,⁹ calculated by subtracting the percentage of homeowning households of people of color from the percentage of homeowning white households. Homeownership has been a significant source of wealth accumulation in the U.S., but many people of color have not been privy to it. Historically, legal barriers to home ownership were removed by the 1862 Homestead Act, guaranteed by the 14th Amendment to the U.S. constitution, and fair housing rights were pledged in the Fair Housing Act (FHA) of 1964. However, discrimination in homeownership continued in the marketplace through white flight, real-estate agents steering races away from each other, and bank redlining and refusing mortgages to minorities while promoting home ownership among the whites (Baradaran, 2019; Florida, 2017; Katznelson, 2005; Taylor, 2019).

These discriminatory practices were commonplace even in some of the most liberal areas, e.g., “… the FHA agreement specified that no properties be sold to African Americans. The builder then constructed individual homes for sale to whites in “Ladera,” a subdivision that still adjoins the Stanford campus” (Rothstein, 2017, p. 12). Many of these practices were justified on economic grounds, e.g., California Real Estate Association warning that “Negro invasion was imminent and that it would result in collapse in property values” (Rothstein, 2017, p. 12) to avoid explicit racial intent. The discriminatory impact of these practices on homeownership of people of color persists today (Kuebler & Rugh, 2013; Rothstein, 2017; Shapiro, 2006; Taylor, 2019). Mayors have authority to enact several remedies. For example, they can apply for and allocate federal and other funds to help reduce inequality in home homeownership, as well as establish task forces to address social and racial housing inequities and foster initiatives that put city housing goals at the forefront of community planning (Barandaran, 2019; Einstein et al., 2017).

Poverty Gap

The third indicator variable of racial inclusion is a racial poverty gap (Shearer et al., 2017).⁷ This is calculated by subtracting the poverty rate (people below the poverty line divided by people in the population) for the whites from the poverty rate of people of color. Though racial inequality in poverty declined in the last 30 years, racial income disparities remain (Baker et al., 2022). The Urban Institute included this as a measure of racial inclusion because occupational segregation persists, with racial minorities tending to be concentrated in jobs with lower levels of stability and with fewer advancement opportunities (Parcel & Mueller, 1983; Smith, 2002). Thus, a lower poverty gap would suggest that leaders have made policy efforts to curb occupational segregation. Although most mayors agree that a racial poverty gap is a problem, they differ on solutions (Smith, 2022). For instance, mayors can enact programs targeted at increasing small businesses owned by minorities and can create programs that improve their access to capital. Some mayors have even adopted a universal basic income program to provide minimum monthly payments to low-income residents or adopted programs that offer financial assistance to minority residents.

Education Gap

The fourth indicator variable of racial inclusion is an education attainment gap (Chapman et al., 2011).¹⁰ This measure is derived by subtracting the percentage of the white population (older than age 25) with a high school degree from the percentage of person of color population with a high school degree, reflecting racial disparities in educational attainment. According to Orfield and Lee (2005, p. 18), more than 60% of minority students attend high poverty high schools compared with 18% of white students. The disparities in funding and access to resources can lead to lower quality education, putting minority students at a disadvantage for future opportunities. Mayors can promote education policies that focus on equality of opportunity. For example, mayors typically have authority to replace an elected school board to raise urban school performance and student achievement. Mayors can also take steps to collaborate with school leaders to keep classroom sizes down, fund programs to ensure students have meals and that their mental health is attended to, raise money, and create cabinets of representatives from school boards, teachers’ unions, and other community leaders to help brainstorm ways to create a city-wide culture that better supports student learning.

Economic Health

An outcome variable in our model, this construct represents the strength of a city’s economy. To mitigate endogeneity, the Urban Institute relied on indicators of economic health that tend to have inclusion-neutral economic growth measures. That is, racial inclusion and economic health covary as exogenous constructs, but not because of endogenous collinearity.

For each observable variable indicating economic health (described below), raw scores were collected and converted by Urban Institute to z-scores, reflecting the relative score of each city on the variable in comparison to the full 274 city sample. Thus, economic health is measured as an index, or average Z-score, across four indicators chosen by the Urban Institute. The z-score was computed by Urban Institute such that a high score represents better economic health.

Employment Growth

The first indicator variable of economic health is employment growth. This measure represents the percent change in the number of people who are employed in the labor force from 10 years prior.¹¹ The Urban Institute included employment growth in the economic health index because employment growth is beneficial for a city’s economy in multiple ways, contributing directly to its Gross Domestic Product (GDP, briefly defined as work hours plus productivity) (Benner & Pastor, 2015; Shearer et al., 2017). Mayors can promote employment growth through workforce development policies. For example, city policies can broaden economic opportunities for urban residents by enhancing demand and job supply (e.g., property tax abatement to promote growth in distressed regions), raising demand for certain groups (e.g., income or tax credits conditional on hiring certain groups), improving the qualifications of workers (e.g., offering free training programs in partnership with local community colleges), and augmenting job attractiveness (e.g., minimum wage increases) (Bartik, 2016).

Unemployment Rate

The second indicator variable of economic health is the unemployment rate. This measure represents the percentage of residents who wish to participate in the labor force but cannot find employment. Federal Reserve Chairwoman, Janet Yellen (2013), stated that the unemployment rate is the “best single indicator of current labor market conditions.” A high unemployment rate has negative consequences for economic health of cities, directly reducing its GDP. It has long been argued that the unemployment rate also captures the degree of job insecurity (see Molotch, 1976) as well as inequalities in employment. The Urban Institute included it as a measure of economic health because when labor is fully employed, economic output cannot grow faster than productivity, unless companies hire more workers. If mayors enact policies that allow wages for low-income workers to drop while keeping unemployment benefits the same, on average, more people will leave their jobs (Krugman, 1994). In this way, employment growth and the unemployment rate are not necessarily two sides of the same coin, but together they offer a more complete picture of a city’s local economy.

Housing Vacancy Rate

The third indicator variable of economic health is a housing vacancy rate, which indicates housing demand and location desirability (Cunningham & Droesch, 2005). It is calculated by dividing the percent of units for sale or rented by the number of sold but unoccupied units. It was chosen by the Urban Institute to measure economic health rather than monthly rent or home sales values as it tends to be the first indicator of a shifting market (Browne, 2018). A high vacancy rates signals poor market conditions and low land-use efficiency. Sometimes, infrastructure growth can exceed actual urban needs, resulting in unutilized capacity. Other times, lack of economic opportunities creates urban shrinkage, resulting in abandoned housing (Haase et al., 2014). Housing vacancies introduce safety issues, as conditions become conducive to crime and reduce a city’s tax base, which can negatively affect the economy (Accordino & Johnson, 2000). Mayors can impact vacancy rates by reducing sale time of vacant properties, demolishing unsuitable buildings, altering eviction policies, funding programs to rehabilitate properties, and streamlining permitting processes.

Median Family Income

The fourth indicator of economic health is median family income, representing a families’ relative prosperity, or economic well-being, within a city, over the past 12 months (Benner & Pastor, 2015; deSouza-Briggs et al., 2015).¹² Median family income is influenced by multiple factors, e.g., occupational structure and industrial mix, educational and gender composition of the workforce, and disability status of a city’s population. According to Manduca (2018, p. 182), “In 1968 … the median family income of African Americans was 57 percent that of whites. In 2016, after almost 50 years of anti-discrimination legislation, attempts to equalize access to education, and cultural change, it was 56 percent.”

This gap persists despite disparity reductions in education attainment (Ryan & Bauman, 2016), life expectancy (Kochanek et al., 2013), and improvement in racial attitudes toward minorities (Bobo et al., 2012). Another reason for the persistence of a racial income gap is the shape of the income distribution itself (Manduca, 2018), as a disproportional shift has unfolded with a drastic increase in the share of economic resources going to the very wealthy combined with income stagnation for rest of the population (Oxfam, 2017). This has disproportionally harmed minorities, who remain overrepresented in the low side of the income distribution (Bayer & Charles, 2018). Mayors can take steps to improve family income, e.g., by increasing minimum wages, enacting guaranteed income programs, improving low-cost city transportation, and lowering other cost of living.

Control Variables

Mayor Race

We included this control variable to estimate the role of gender more accurately, as a mayor’s race could influence the priority level placed on inclusion.¹³ A research assistant blind to the hypotheses coded mayor’s race to align with Urban institute’s definition of residents of color, as defined earlier in this article. Pictures and biographies available online (e.g., city websites, news articles, obituaries, LinkedIn) were used to code mayor race. Of the 971 unique mayors in the dataset, 92 were coded as a person of color (9.5%).

Term Length

Prior leadership research, as reported in Bernerth and Aguinis (2016), commonly controls for tenure-related variables to account for differences in accumulated knowledge and experience. Moreover, because mayoral electoral cycles vary by city, there are differences across mayors relating to the amount of time each had to enact policy changes before the measurement time point. For example, if a mayor was elected in 2010, for the 2013 inclusion and economic health numbers, s/he would have had 3 years to enact changes that might impact those measures. Thus, we control for the mayor’s term length prior to the year of measurement by subtracting the year of measurement from the first year in office.¹⁴

Political Affiliation

Because political affiliation differentially affects policy decisions, we controlled for the mayor’s political party. Specifically, we coded the mayor’s political affiliation as democrat, republican, or non-partisan. Of the 971 unique mayors in the dataset, 48% were democrat, 20% republican, and 32% were non-partisan.

City Population

Smaller cities tend to be more inclusive on average and the greater the city population, typically, the more complex it becomes for leadership to navigate issues related both to racial inclusion and economic health (Poethig et al., 2018). Moreover, the greater a city’s population, the faster it can experience economic growth and offer economies of scale (Simon, 1998). For these reasons, we include city population from the U.S. Census Bureau as a covariate.

Percentage of People of Color

The percent of a city’s population who self-identified as persons of color impacts inclusion and economic health. Mayors with more diverse cities manage more complexities in achieving inclusion and equality of economic prosperity (Florida, 2017).

Analytical Procedures

Given our interest is examining how gender, a categorical variable, relates to quantitative outcomes, we tested our model with analysis of covariance (ANCOVA) using a linear mixed effect model (LMEM) with centered regressors for gender (female 0.5; male  − 0.5).¹⁵ LMEM is a form of hierarchical linear modeling (HLM) that uses restricted maximum likelihood to account for non-independence (e.g., repeated terms served by the same mayor, repeated data for the same city across time, and macro-year effects) by parsing out random effects to obtain the least-biased estimates of the fixed effects (Brauer & Curtin, 2018). The data included clusters of cities nested within states, mayors nested by name to account for non-independence if the same mayor served at multiple data points (e.g., 2013 and 2016), and city-state observations nested by year. This parses out random variance in inclusion and economic health related to city and state, as well as macro-year factors, reducing error in the individual-level (gender) fixed effect estimate.

For example, cities might differ on broad measures of racial inclusion and economic health (e.g., government funding, availability of federal programs), and inclusion of a random intercept for city-state makes it explicit that these differences across levels could impact the criteria. Moreover, including a random intercept for mayor name accounts for non-independence in outcome measures if the same mayor served in multiple terms, rather than treating the same mayor at two different time points as separate individuals. Including a random intercept for year accounts for macro-level influences that impact economic health and inclusion of all cities (e.g., nation-wide recession). Said differently, a model without random intercepts (i.e., a non-nested model) would assume that racial inclusion and economic health across cities, years, and mayors would be expected to have the same variability; this assumption is likely untenable.

For the mediation analysis, we followed Baron and Kenny (1986) procedures. We first tested the link between mayor gender and economic health. Then, we examined the relation between mayor gender and racial inclusion. We followed this with the full mediation model in which racial inclusion is added as predictor variable for economic health. Preacher and Hayes (2004) advocated bootstrapping to compute a confidence interval for indirect effects. Thus, we used the Imai et al. (2010) mediation package in R, which relies on a quasi-Bayesian approximation using 1000 simulations, to estimate confidence intervals and significance levels for the indirect effect, total effect, and proportion of effect mediated.

Results

Confirmatory Factor Analysis (CFA) of the Measurement Model

We first verified whether the indicator variables loaded onto the constructs of interest (racial inclusion; economic health) to ascertain the appropriateness of using the index measure. For construct validation of the comparative racial inclusion index, we ran a CFA grouped by year (DiStefano & Hess, 2005) to test if the four inclusion variables indeed manifest racial inclusion. Goodness of fit indices indicated data fit the measurement model.¹⁶ Table 1 presents the mean, standard deviation, and pair-wise correlations for each racial inclusion indicator variable.

Table 1 Means, standard deviations, and pair-wise correlations for racial inclusion indicator variables

For construct validation of the economic health index, we ran a CFA grouped by year to examine if the four economic health variables indeed manifest economic health. Goodness of fit indices indicated data fit this measurement model.¹⁷ Table 2 shows the mean, standard deviation, and pair-wise correlations for each economic health indicator variable.

Table 2 Means, standard deviations, and pair-wise correlations for economic health indicator variables

Preliminary Results

Table 3 presents descriptive statistics and bi-variate correlations. The three focal variables—mayor gender, racial inclusion, and economic health—wer significantly correlated. The direction of the correlations suggests that cities with female mayors, on average, were associated with greater racial inclusion and better economic health. The bi-variate correlations likewise supported our rationale for inclusion of the control variables, as mayor race, mayor political affiliation, mayor term length, city population, and person of color percentage were significantly related to economic health, and all but mayor race were significantly related to racial inclusion. As the covariates are not focal variables (i.e., included to parse out noise only), we do not speculate on the direction of those correlations.

Table 3 Means, standard deviations, and bi-variate correlations

Test of Hypotheses

Following ANCOVA procedures (Keppel, 1991), to test the unique relationship between mayor gender (dichotomous variable) and the outcome (continuous) variables, we parsed out the covariance between control and outcome variables. The mediation procedures outlined by Baron and Kenny (1986) examine  four key conditions: (1) gender (x) effect on economic health (y), (2) gender (x) effect on racial inclusion (m), (3) racial inclusion (m) effect on economic health (y) when controlling for the effect of gender (x) on economic health (y), and (4) the effect of gender (x) on economic health (y) disappears or becomes significantly smaller when statistically controlling for the effect racial inclusion (m) on economic health (y) (i.e., significance of the indirect effect).

Table 4 reports the ANCOVA model results testing all four conditions of mediation with the full dataset. Table 5 reports the ANCOVA model testing all four conditions of mediation with a constrained dataset in which we only included mayors who were elected to office at least four years prior to measurement of the criterion variables (n = 536).¹⁸ Of the control variables, city population significantly related to economic health in both models such that larger cities were associated with a higher economic health index. H1 was supported in both the full and constrained models, as mayor gender was significantly related to economic health such that cities with female mayors were associated with better economic health than cities with male mayors.

Table 4 Full dataset model results

Table 5 Constrained model results (elected as mayor at least four years prior to measurement)

To test the second condition of mediation, whether mayor gender relates to racial inclusion, we included median family income as a covariate to account for the possibility that cities with higher racial inclusion index scores were in places where people of color were priced-out instead of cities where they are present in reasonable numbers and included in civil and economic life.¹⁹ City population, person of color percentage, and median family income significantly related to racial inclusion such that more populated cities and more diverse cities had worse racial inclusion scores and cities with higher family incomes had better racial inclusion outcomes. Supporting H2, mayor gender significantly related to racial inclusion. Cities with female mayors were associated with better racial inclusion than cities with male mayors.

For the third and fourth conditions of mediation, city population and person of color percent remained as significant control variables related to economic health. Supporting the third condition of mediation, racial inclusion was significantly and positively related to economic health. Supporting the fourth condition of mediation, the gender association with economic health became marginally significant when racial inclusion index was added as a predictor in the model. Together, both the full and constrained dataset appear consistent with a mediation model.

Following Tingley et al. (2014) mediation procedures in R, we assessed the significance of the indirect effect of mayor gender on economic health through racial inclusion to test Hypothesis 3. As reported in Table 6, in both the full and constrained dataset, the indirect gender effect was significant, the total gender effect was significant, the direct gender effect became nonsignificant when the mediator was controlled for, and the proportion of the gender effect mediated by the racial inclusion index was significant. These findings indicate that the data are consistent with the hypothesized mediation model, supporting H3. These relationships, as estimated from the constrained dataset, are illustrated in Fig. 3.

Table 6 Mediation results

Fig. 3

Illustration of the mediation findings from the constrained model. This figure illustrates the results of the mediation analysis with the fixed effect estimate presented on the lines between constructs with a p value indicator as follows: ***p < 0.001, *p < 0.05

Because these panel data are correlational, to mitigate reverse causality and alternative explanations, we performed several robustness checks by cross-lagging the data backward. First, we confirmed that mayor gender in year t was unrelated to economic health in year t–1 (p = 0.746) and unrelated to racial inclusion in year t–1 (p = 0.222) (e.g., mayor gender in 2016 was unrelated to economic health and racial inclusion in 2013).²⁰ Next, using a generalized linear model with the binomial family and logit link function and controlling for city-state and year fixed effects, we confirmed that economic health in year t–1 was not a significant predictor of mayor gender in year t (p = 0.842) nor was racial inclusion in year t–1 a significant predictor of mayor gender in year t (p = 0.363). This mitigates reverse causality by showing that a more racially inclusive or economically healthier city was no more likely to elect a female mayor compared with other cities.

Discussion

We built on FLA literature and integrated it with ethics of care to conceptualize how women leaders might achieve urban revitalization. FLA findings guided our first hypothesis that female mayors might be more effective than male mayors in managing the urban revitalization crisis. We next proposed and empirically tested the mediating role of ethics of care leadership in the relationship between women-led cities and urban revitalization. Ethics of care leadership parts ways with traditional think-business think-inclusion expectations, i.e., more money leads to more prospects, and benefits will be naturally shared equally. It shifts the focus to a leader’s ethics as the principal guide for action in revitalizing urban areas. Ethics of care leadership views the interests of those caring and those cared-for as interwoven, and it nurtures inclusion among residents as a path toward achieving economic health. Results from 272 U.S. cities over nearly four decades revealed that cities with female mayors were associated with greater racial inclusion and better economic health than cities with male mayors. Mediation analysis supported ethics of care as a pathway to a FLA, as better economic health in women-led cities was explained by their association with greater racial inclusion. These findings make several contributions.

First, scholars have been calling for leadership research to take on a greater role in addressing social issues (Walsh et al., 2003). We extend a FLA to urban revitalization, and, in doing so, contribute to emerging research that examines how leadership effectiveness can transcend financial contributions to organizations, extending significant societal impacts as well (Chang et al., 2022; Sergent & Stajkovic, 2020). Preferences for women leaders during an organizational crisis have been documented (Ryan et al., 2016), but research on actual leadership effectiveness throughout a chronic crisis that unfolds over time is rarer (see Post et al., 2019). This distinction, coupled with empirical tests of real-world data, differentiates between a focus on expectations of leadership appointments and actual leadership outcomes.

Second, our backward cross-lagged robustness check indicated that more racially inclusive cities and healthier cities were no more likely to elect female mayors in the future than less racially inclusive or less economically healthy cities. This finding helps mitigate reverse causality. The time lagged design overcomes a limitation of prior research that has relied on a single period to investigate a FLA. For example, without time-variant data, it is difficult to empirically disentangle if female leaders are associated with more effective outcomes because of their leadership style (e.g., ethics of care), or if citizens (employees) in female-led cities (organizations) differ systematically in some way that contributes to differences in outcomes.

Third, our results have reasonable ecological validity regarding urban revitalization of cities. Because our data were not experimental, however, causality cannot be inferred. Typically, researchers do not manipulate inhumanities, and therefore, racial inclusion/exclusion experiments would be limited to hypothetical scenarios. Verification of causal links between urban revitalization and socially situated outcomes will require converging evidence from divergent methods. We call for continued research using varied methods to build evidence toward causality between leader gender and leadership effectiveness.

The data show that economic health in the U.S. has been uneven—both between cities and within them (Funders Network for Livable Communities, 2017). Because many cities have seen widening racial inequalities despite economic gains, city leaders are increasingly linking inclusion goals to economic health. Economic growth does not automatically lead to racial inclusion. For example, Poethig et al. (2018) analyzed economic recovery of four cities after the Great Recession of 2007–2009 and found that adopting a shared vision and collaborating across sectors helped promote economic recovery, but these actions were not enough to promote racial inclusion. Leaders who furthered inclusion devoted effort to elevating the voice of minority residents, enacted policies that removed barriers, and crafted narratives that bound together inclusion and economic growth. By linking racial inclusion to economic health, more city residents can contribute to the economy and share in the benefits.

Limitations and Future Research

First, because our data are correlational/archival, no experimental causality can be claimed, only patterns of associations. Though results revealed a pathway from mayor gender to economic health via racial inclusion, this study does not reveal specific actions women mayors took that resulted in female-led cities being associated with greater racial inclusion than male-led cities. We theorized that divergences among cities in their degree of racial inclusion manifest because of differences in the embodiment of ethics of care leadership by mayors. Future research can design randomized experiments to causally tease out links between leader gender, ethics of care, and urban revitalization.

For instance, an experiment could be conducted that requires participants to act as a leader in a hypothetical scenario aimed at improving economic growth of a city. Participants could be presented with choices of policies to enact, where each policy costs a finite number of resources and participants make choices among policies by spending resources. This would force tradeoffs whereby participants have to allocate limited resources to policies they feel represent the best pathway toward city development. Each policy would have pros and cons described; some policies could have a clear undertone of ethics of care leadership—promoting goals related to racial inclusion and emphasizing the downstream impact on local residents. Other policies could have undertones related to business growth. Participants could be asked to explain their reasoning after makingchoices. Participants' selections could be analyzed to assess if female participants were more likely than males to choose options associated with ethics of care and to what cost they were willing to go to do so, and similar comparisons could be made for options related to economic growth. Perhaps an interesting comparison would be to examine changes in ethics of care by gender holding preferences for business constant. This would further illuminate our findings by revealing whether women leaders care more about communities because they care less for business or if they care equally about business but care more for others on top of it. Qualitative analysis could be performed on the open-ended explanations to build nuance in the theory to further explain why females appear more likely to enact ethics of care leadership.

Second, although care is a virtue, it can come with drawbacks if used in excess (Smith, 1998). An extreme emphasis on care could undermine the volition of those cared-for, especially if the care receiver is provided with more care than desired or needed (Baier, 1994; Lloyd 1984). There are degrees of care (Slote, 2007), and there is a fine psychological line between overindulgent care and a perception of pity. Thus, effective leaders strike a balance between inadvertently conveying pity rather than respect for the dignity of those receiving the care. This is why the framework was labeled ethics of care, rather than just care (Held, 2006, p. 11).

Third, the dataset was constrained by the parameters used by the Urban Institute, such as inclusion of only U.S. cities with populations exceeding 100,000 in 2016, and time periods spanning 10, 13, and 3-year gaps. Thus, future research is needed to better understand if the findings generalize beyond U.S. borders, outside of populated areas, and to other time periods. Inclusion of 1980 and 1990 in the analysis somewhat mitigates a generalizability concern to smaller cities, as many cities in the dataset were less populous at that time. Nesting the data by year and controlling for population, we continued to find evidence of a FLA, which suggests that FLA holds regardless of city-size. We included covariates and city-state fixed effects to control for mitigating factors, such as existing racial diversity, population, mayor race, and political affiliation, but we did not speculate on the direction of those relationships.

Future research could examine whether there are moderating factors that accentuate or attenuate the mediating effect of racial inclusion as it relates to a FLA. For example, length of time living in the city (or organizational tenure) might moderate a FLA, such that leaders who identify more with their communities and bond with members through lived experiences might be more likely to focus on ethics of care when leading through a crisis. Moreover, future research can investigate the origins of gender differences in ethics of care leadership to more precisely discern whether differences arise from varying degrees of value placed on care as a moral virtue, as we suggested, or whether these differences relate more to societal pressure on women to conform to gender stereotypes or to a greater likelihood of selecting women who conform to existing norms for leadership roles.

Fourth, this study examined the role of leader gender on outcomes, controlling for race and political orientation. It did not take an intersectional approach by conceptualizing and testing an interactive model among these variables. This was an a priori design choice given the deductive nature of this study. Specifically, we sought to examine a FLA in the context of urban revitalization by conceptualizing and testing ethics of care leadership as a potential mechanism through which women leaders associate with an advantage. It was not an event-focused study, in which race might be more of a main factor considered. We encourage future research to build upon the theoretical foundation of ethics of care leadership, as described herein, by inorpating an intersectional perspective. Such research should aim to provide detailed, theory-driven explanations of why and how gender, race, and other identity-related variables interact in relation to ethics of care leadership.

Conclusion

Most people in the U.S. live in cities. As cities grow, so do disparities in racial inclusion. Policies like the Fair Housing Act were passed to increase inclusion but declines in racial segregation have not unfolded as expected (Krysan & Crowder, 2017). Revitalizing urban environments is a societal challenge with multiple implications. Bridging leadership research on a FLA and ethics of care theory in moral psychology, we proposed that mayoral leadership based on ethics of care balances the focus on racial inclusion and economic growth in the cities as a sustainable direction going forward. On the basis of data from 272 U.S. cities at five points in time, and multiple measures of key constructs, we found that cities with women mayors were associated with greater racial inclusion and better economic health, and mediation analysis revealed that racial inclusion was one mechanism through which women mayors impacted economic health. These results from a setting with real-world complexity indicate that the urban crisis is a challenge that can be meaningfully addressed by leadership and ethics research.