Abstract
One of the primary ways cities signal the importance of diversity is to hire a chief diversity officer (CDO). Despite the importance of this position, we know remarkably little about the prevalence and predictors of CDO adoption in local governments. Using original data from the 250 largest cities in the United States, this article demonstrates that about one in four cities has hired a CDO. Further, the decision to hire a CDO is driven more by constituent demand than by city structure or resources. The article concludes by discussing the future of diversity management in America’s cities.
Over the past few decades, the government workforce has become more diverse (Riccucci 2001), but minorities continue to express relatively low levels of trust in government (Howell and Perry 2004; Rahn and Rudolph 2005). To address these seemingly contradictory trends, managing diversity in a proactive manner has become a more salient goal for many local governments (Becker 2017). Although cities have a number of potential policy solutions to address diversity, one of the most frequently used strategies is to hire a diversity advocate (Hur and Strickland 2012) such as a chief diversity officer (CDO). Hiring a CDO is important because of the substantive work of the CDO and because it reflects a conscious choice to make a meaningful capital investment in diversity management. Further, hiring a CDO reflects a commitment to increased oversight of managers and street-level bureaucrats alike.
Anecdotal evidence suggests that CDOs are frequently used and are becoming more ubiquitous in local government management. However, scholars know surprisingly little about the decision to hire a CDO. The literature provides no evidence about how many cities have CDOs, when they adopted them, or why. This article answers those questions using a unique data set of CDOs in the largest American cities. Answering these questions is important for academics who wish to understand the causes of policy innovation and diversity management as well as for practitioners who may want to benchmark their municipality’s diversity-related initiatives off of other similarly situated governments.
Managing Diversity
While articles examining CDOs in local governments are few and far between, there is considerably more work on the causes and effects of managing diversity in general (Battaglio and Hall 2018; Alkadry, Blessett, and Patterson 2017). Rather than taking a passive approach to diversity, managing diversity “refers to the ability of top management to develop strategies as well as programs and policies to manage and accommodate diversity in their workplaces” (Riccucci 2002, 3). One model of diversity management suggests that proper diversity management consists of three primary components: recruitment and outreach, valuing differences, and pragmatic policies and programs (Pitts 2006). Within these broad components, a number of specific, tractable strategies can be adopted to manage diversity across race, gender, age, ability, and sexuality. Some of the most common of these strategies include increasing promotion and advancement opportunities for diverse people, eliminating harassment, developing leave and flex policies that reflect people’s unique situations, ensuring that the organization and its employees adhere to policy and law as they apply to human resources, creating mentorship programs, and partner benefits for employees of all sexual orientations (Ewoh 2013; Riccucci 2002). A transformational leader (Kearney and Gebhert 2009) such as a CDO may facilitate all of these strategies and thus may improve diversity management within the organization. The presence of an effective CDO may therefore be an outward sign of an organization that places a premium of managing diversity.
A number of scholars have also examined the effectiveness of diversity management in creating favorable outcomes for public organizations. For example, an organizational commitment to diversity management can increase both job satisfaction and work group performance (Pitts 2009). Similarly, fostering an inclusive work climate can reduce discrimination and workplace bullying (Andrews and Ashworth 2014). Of course, this is not to suggest that all employees perceive organizational climate similarly; employee perceptions of a diverse climate within the organization vary based on their social identities and perceptions of procedural justice (Oberfield 2015).
Insights about diversity management can also be found within the vast literature on representative bureaucracy. Simply put, the representative bureaucracy literature argues bureaucracies should resemble the people they serve. A more representative bureaucracy isn’t simply about descriptive representation as an end unto itself, however. Representative Bureaucracy can help a bureaucracy move from passive to active representation and can increase governmental legitimacy (Riccucci and Van Ryzin 2017). Considering that CDO adoption should signal a government that places a premium on diversity management, it stands to reason that local governments with CDOs should, on average, do a better job hiring and retaining a diverse workforce and should therefore be more representative of the people they serve.
CDO Pacesetters
The CDO position is not unique to local governments. In fact, there is anecdotal evidence that others sectors, most notably private businesses and higher education, have employed CDOs for over 20 years. Roughly 20 percent of Fortune 1000 companies have a CDO or similar position (Rayome 2016). These include tech giants such as Pinterest, Salesforce, Google, Facebook, Microsoft, Airbnb, Dropbox, Twitter, and Apple (Templeton 2017). This aggressive adoption of the CDO model seems to pay dividends for the organizations that choose to follow it. For example, companies with commitments to diversity tend to be more innovative and more profitable than companies that are less committed to diversity (Hewlett, Marshall, and Sherbin 2013). CDOs facilitate these favorable outcomes by supporting customer collaboration and encouraging diverse perspectives in product development (Guillory 2017).
There has also been a rise in CDO positions on America’s college campuses. For example, in the eighteen-month period immediately prior to 2016, ninety American colleges and universities hired CDOs (Frum 2016). Despite ambiguous roles and relatively little managerial guidance (Brown 2017), CDOs continue to be more and more prevent as higher education sees the benefits of diversity, equity, and inclusion. Some speculate that this rapid and relatively recent rise in CDO adoption is driven by environmental factors such as increased racial tensions on campus (Frum 2016) and pressure from student activists (Brown 2017).
Hypotheses and Measurement
To learn more about the prevalence and correlates of a city’s decision to adopt a CDO, this article relies on an original data collection effort of the 250 largest cities in the United States (U.S. Census Bureau 2018). The data collection catalogued whether a city had a CDO and when the first CDO was hired. About a quarter of the time (23 percent), news articles (either through a simple Google search or from newspaper archives) reported the results of a hire and noted when the first CDO was hired. Other times information was available on the city’s website. In the remainder of the cases, it was necessary to call the city’s human resources department to determine whether the city had a CDO, and if so, what their official title was and when they were hired. While CDO is the most used title in municipal government, forty-four other titles were used in large cities such as ethics and diversity officer and civil and human rights director.
While we can find no extant study that addresses this topic, the literature suggests three potential explanations for which cities might choose to adopt CDOs: resources, constituent demand, and city structure. Those are described below.
Resources
It stands to reason that resources should be a driver of CDO adoption. After all, municipalities with more financial resources have the fiscal flexibility to engage in a host of activities (Aldrich 1979). In the context of diversity management, Fernandez and Rainey (2006) and Pitts (2007) suggest that resource-rich organizations are the most likely to engage in diversity management. It is, therefore, likely that cities with more resources may be more likely to invest in hiring a CDO—and undergoing the necessary changes that decision implies. This is termed the resources hypothesis. While this is still the most likely possibility, it is notable that a similar measure (per pupil revenue) had no discernable effect on the adoption of diversity management programs in Texas school districts (Pitts et al. 2010). To operationalize the resource hypothesis, the analysis to follow relies on the Annual Survey of State and Local Government Finance (2017), which provides budget information for every municipality in America.
Constituent Demand
The second hypothesis builds on theories of representative bureaucracy and constituent demand. The expectation is that local governments will respond to the anticipated needs of their constituents in two specific ways. First, governments whose citizens are made up of more diverse populations will be more likely to respond by adopting a host of policies to manage diversity—including creating a CDO. This follows naturally from the well-established literature on representative bureaucracy (e.g., Bradbury and Kellough 2011; Riccucci, Van Ryzin, and Jackson 2018) and diversity management (Hur and Strickland 2012). Similarly, it is recognized that administrators operate in a political environment, and the decision to hire a CDO is no different. Therefore, the hypothesis holds that the decision to adopt a CDO is more likely to be welcomed in a city with a more liberal population. After all, conservatives may dismiss CDOs as mere “virtue signaling” (Leef 2018). Together, these variables test the constituent demand hypothesis.
The constituent demand hypothesis necessitated two additional variables—one which measured ideology and one which measured diversity of the city. Citizen ideology has traditionally been difficult to estimate at the local level because of the difficulty in generating large sample in relatively small geographic areas. Fortunately for scholars of local politics and administration, new measures get around this problem by applying multilevel regression and prostratification to aggregate public opinion polls into a stable, robust indicator of subnational public opinion that is available across cities (Tausanovitch and Warshaw 2013; 2015). Tausanovitch and Warshaw’s ideology scores are estimated across a traditional liberal–conservative ideological spectrum where higher numbers indicate a more conservative populace. We apply these scores as a proxy for generalized political ideology.
To measure diversity, this analysis relies on the U.S. Census Bureau (2018) to estimate the proportion of the city’s population that is Latino, black, white, and Asian. Following Sullivan (1973; see also Hero and Tolbert 1996), these data are then translated into a single diversity index using the following formula:
City Structure
The voluminous literature on city structure suggests that council–manager governments are more likely to adopt a host of innovative policies than are their counterparts in mayor–council cities (Feiock, Steinacker, and Park 2009; Krebs and Pelissero 2010; Nelson and Svara 2012). City managers are much more likely to have graduate education in public administration than their elected counterparts. As a result, they are more likely to be attuned to best practices in the field, such as increasing emphasis on diversity management and the benefits of hiring a CDO. Given these findings, it is expected that council–manager governments are more professionalized, are more likely to be informed by current practices in diversity management, and therefore more likely to adopt a CDO. This is termed the city structure hypothesis. These data are collected from the Census of Governments (2017).
There are, of course, a host of other potentially important variables that could influence the adoption of CDOs. These include, but are not limited to, the incidents of lawsuits based on discrimination and the racial/gender mix of the city council. Nonetheless, it is theoretically and empirically advantageous to include only these four variables that have clear theoretical justification and are reliable and empirically verifiable. This parsimonious model is also preferable, given the relatively small sample size and likely multicollinearity between other potential independent variables.
Descriptive statistics for all variables used in this analysis are available in Supplemental Table 1.
Results
This analysis begins by reviewing the most basic, but perhaps the most important, descriptive statistic in this study—the percentage of cities that have hired a CDO. Even in this sample of large cities, only about one in four (24 percent) of the cities in our study have hired a CDO. CDO adoption is not even universal among the ten largest cities (three of which have never had a CDO in the city’s history). Clearly, hiring a CDO as a way to manage diversity is far from a ubiquitous strategy in America’s largest municipalities. It stands to reason that CDO adoption would be even less prevalent in small- and medium-sized cities.
Figure 1 presents the date that a CDO was initially hired in bars (presented as the number of cities that hired their first CDO in a given year), and the cumulative percent presented in the dotted line. This figure reinforces the conclusion that CDO adoption is far from ubiquitous. In addition, this figure suggests that the drive for CDO adoption is relatively in the vast majority of cases. In fact, more than 50 percent of all initial CDO adoptions in America’s biggest cities have occurred since 2006! While the first cities adopted CDOs in the 1950s, it took half of a century before CDO adoption could be described as anything other than an outlier. Altogether, this figure suggests that the public sector is surprisingly late in CDO adoption, but recent years have seen a precipitous increase in CDO adoption. If a large city does not have a CDO today, they will have a lot of company, but if current trends continue, a city without a CDO will soon find itself as the outlier.

CDO adoption 1950–2017. Source: Authors’ data from 250 largest cities in America.
Table 1 presents the results of a logistic regression where the dependent variable represents whether a city has adopted a CDO (1) or has never had a CDO (0). The model includes independent variables for city diversity and political ideology (both testing the constituent demand hypothesis), the city budget (testing the resources hypothesis), and whether the city has a council–manager form of government (1) or not (0) (testing the city structure hypothesis).
Modeling Chief Diversity Officer Adoption in America’s Largest Cities.
Note: Estimates are from a logistic regression model. Standard errors are clustered on the city.
*p < .05.
**p < .01.
As Table 1 indicates, there is strong support for the constituent demand hypothesis. Cities whose citizens are more politically liberal and who are more diverse are much more likely to adopt a CDO than are cities with comparatively more conservative and less diverse populations. In contrast, the analysis provides evidence to support the resources or the city structure hypotheses. It appears that cities with more resources are no more likely to adopt a CDO than those with comparatively fewer resources. It is important to remember that our sample only includes relatively large cities, so it is possible that the effects of resources are not apparent in this data set but might be in much smaller cities. Further, the null effects of city structure contribute to a large literature on city structure (Carr 2015), and results align with some other studies that find adoption of council–manager government does not necessarily lead to more innovative policies.
These findings are certainly instructive, but logistic regression results are notoriously difficult to interpret in substantively meaningful ways. For example, it is difficult to understand what a logistic regression coefficient of 1.69 means in practice. To better understand the substantive effect of constituent demand on CDO adoption, it is therefore necessary to estimate the predicted probabilities for the statistically significant variables on the probability of adopting a CDO (see Supplemental Figure 1). These estimates are generated using the “margins” command in Stata (version 15; Williams 2012).
The results of this additional analysis in Supplemental Figure 1 reinforce the notion that cities with more liberal citizens are much more likely to hire a CDO than those with more conservative citizens, on average. In fact, controlling for other factors in the model, the most liberal city in the data set has more than a .60 probability of having a CDO, whereas the probability of having a CDO in the most conservative city in the data set is in the single digits.
The results regarding the substantive importance of diversity on CDO adoption tell a similar story. More diverse cities are much more likely to have a CDO than less diverse cities. Specifically, the least diverse city in the data set has less than a .10 probability of having a CDO versus a probability of just less than .50 for the most diverse city in the data set. Clearly, constituent demand and city ideology drive CDO adoption in ways that are substantively important. Supplemental Figure 1 presents all of these results in graphical form.
Discussion and Conclusion
CDOs seem to be growing in number and importance in local governments (Kimbrough 2017). Despite their presumed importance, there is scant empirical evidence about how many CDOs exist or why certain cities choose to adopt CDOs and others do not. This lack of knowledge means that city administrators cannot determine whether adopting a CDO is a “must-do” for a city of their size and profile or whether hiring a CDO would put them on the forefront of managing for diversity. Scholars of diversity, policy adoption, and local government, more generally, are similarly left in the dark about the decision to hire these important public managers.
After collecting and analyzing an original data set from the 250 largest cities in the country, this article concludes that CDO is neither rare nor ubiquitous in local government. Just over one in every four of America’s 250 largest cities has hired a CDO. That number is increasing rapidly, and if current trends continue, it will be just a few years before CDO adoption reaches one half of the largest cities in the United States.
The analysis also finds that constituency pressure rather than resources or city structure explains local government CDO adoption. Cities with a more liberal population and a more diverse population are the most likely to adopt CDOs. This suggests that hiring a CDO may be a measure of active representation, where bureaucracies respond to the needs and desires of their communities. The lack of support for the resources hypothesis is consistent with Pitts et al. (2010) who found that schools with greater resources were no more likely to implement diversity programs. It seems that diversity, in general, is predicated more on constituent demand than slack resources.
While further support for active representation in local bureaucracies may be seen as a positive, it raises questions about diversity management in cities where the citizens are less likely to advocate for CDOs and other similar diversity-related initiatives. This gap may be particularly stark in rural governments that already face substantial management challenges (Helpap 2019). If diversity management truly is necessary for a well-functioning bureaucracy, then scholars should seek ways to “close the gap” between more liberal, diverse cities and those with a more conservative and less diverse citizenry.
While these findings are instructive for both practitioners and academics, they are, of course, the first rather than the last words on this subject. This article provides no empirical evidence from smaller cities. If resources do not matter, is it possible that smaller cities with more diverse and liberal populations might adopt CDOs before their larger and less diverse counterparts. If city size is an important predictor, it would also imply that large cities may be better choices for both employment and residency for diverse populations and those who wish to live and work in diverse environments.
This analysis also provides no evidence about the content of the work of CDOs in these cities. As the CDO of Buffalo, New York, said, to be successful, “the role of the CDO ‘has to be genuine’ and ‘can’t just be a figurehead’” (Kimbrough 2017, 63). Recent work by Rimes and Long (2019) augments this conclusion. While we are treating CDO adoption as ultimately a “good thing” and a reflection of a city that is attempting to manage diversity successfully and proactively, future scholars should test the veracity of this assumption. As one of the first scholarly articles about CDOs in the public sector, this article should be seen as the beginning, rather than the conclusion of a discipline-wide research agenda about the causes and effects of hiring a CDO.
Supplemental Material
Supplement_Material - Diversity Management in Action: Chief Diversity Officer Adoption in America’s Cities
Supplement_Material for Diversity Management in Action: Chief Diversity Officer Adoption in America’s Cities by Christopher A. Cooper and John D. Gerlach in State and Local Government Review
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
