Abstract
This article “goes local” to investigate the representation of women on city councils, seeking explanations for the variation in women’s descriptive representation at the municipal level. Using the State of New Jersey (NJ) as a case, it is asked: what explains why women fare electorally better in some NJ municipalities than in others? More specifically, what explains “blanks”—or councils on which women are absent—in women’s representation in local politics? It is demonstrated that council size is a significant predictor of women’s presence or absence, but not percentage representation, on city councils.
In 2017, two of America’s largest cities made history in electing women mayors. In New Orleans, two women of color competed in the November 2017 runoff, ensuring that the city would elect its first woman—and first woman of color—mayor in its history. In Seattle, another all-female runoff election in November 2017 resulted in the first woman mayor of the city in over ninety years. These contests attracted nationwide attention, not only because these women fill roles that have long been occupied by men but also because victories by women mayoral contenders of major cities continue to be unusual occurrences. What has garnered less attention in both cities is the presence of women on their city councils. Seattle elected a majority-female council in 2015, and women continue to outnumber men in 2018. The New Orleans City Council was majority female from 2008 to 2017.
Cases such as these, where women hold a majority of elected offices at the municipal level, are so rare that they should garner considerable interest, but few citizens, scholars, or advocates have taken note. Likewise, despite considerable scholarly attention to women’s low levels of representation in state and federal offices in the United States (e.g., Sanbonmatsu 2006; Carroll and Sanbonmatsu 2013b; Dittmar, Sanbonmatsu, and Carroll 2018), little scholarship has investigated the underrepresentation of women in municipal offices and sought to identify its causes and effects—whether symbolically or on the substantive work of municipal governance (for exceptions, see Smith, Reingold, and Owens 2012; Holman 2015; Crowder-Meyer, Gadarian, and Trounstine 2015). A recent article in State and Local Government Review explicitly calls for more research on women’s representation in the many local political bodies around the country (Holman 2017).
The tendency for women’s underrepresentation in local politics to fly under the radar might be due in part to the widespread perception that women are doing relatively better the lower you go in the political hierarchy (Darcy, Welch, and Clark 1994, 33; Crowder-Meyer 2013, 399)—a truism which has, however, recently been challenged both theoretically and empirically (Kjaer 2019). For example, while women remain underrepresented at all levels of American politics, the proportion of women as mayors in cities over 30,000 (21.8 percent) is lower than the proportion of women as state legislators (25.4 percent; Center for American Women and Politics [CAWP] 2018).
This article “goes local” to investigate the representation of women on city councils, seeking explanations for the variation in women’s descriptive representation at the municipal level. Using the State of New Jersey (NJ) as a case, the question asked is: what explains why women fare electorally better in some NJ municipalities than in others? More specifically, the focus in this article is on what should be denoted “blanks” (councils where no women are present) in women’s representation in local politics (see Dahlerup and Haavio-Mannila 1985; Dahlerup 1988, 283). For instance, while most Americans are familiar with Atlantic City as a tourist destination and East Coast hub of casino culture, few people likely know that its municipal government consists of nine council members, all men. The mayor, who is the leader of the municipality’s executive branch, is also male. And while Atlantic City might be an outlier among the 36,011 general purpose municipalities and townships in the United States (Leland and Whisman 2014, 416) in terms of the number of restaurants, bars, and casinos, it is not an isolated case when it comes to its all-male city council. In NJ alone, 135 localities exist where only men served as council members at the municipal level in 2014.
After explaining the reasons for selecting NJ as the case for examining women representation, several hypotheses are presented regarding the variation in the presence of female council members statewide. These hypotheses are then tested in the empirical analyses, and ideas are explored regarding how it might be possible to “fill the blanks” by electing at least one woman to councils where none serve. Finally, how changing the number of council seats can both help to eliminate blanks and spur opposition to this type of electoral engineering is discussed.
Women in NJ Local Politics
Research on women and politics at the local level in the United States has been scarce (for exceptions, see Smith 2014; Crowder-Meyer, Gadarian, and Trounstine 2015; Holman 2015; Carroll and Sanbonmatsu 2013a), and existing studies have had limitations. The three most well-known and extensive studies include only large municipalities, thus limiting the ability to investigate the influence of municipality size and urbanization on women’s representation (Welch and Karnig 1979; Bullock and MacManus 1991; Smith, Reingold, and Owens 2012). Trounstine and Valdini’s (2008) study is distinct in its breadth and depth including 7,174 municipalities with more than 2,500 residents. However, their reliance on survey data means that they miss some nonrespondents in presenting the most complete picture possible (Trounstine and Valdini 2008). Including municipalities of varying sizes, not just large ones, is essential to fully understanding women’s representation as a function of municipality characteristics. Recognizing the methodological hurdles to including or sampling all municipalities nationwide (Holman 2017, 292), the focus in this article is on one state—NJ—that offers considerable variation in municipalities’ size and characteristics for the analyses of women’s local representation.
There are multiple reasons why NJ is particularly well-suited for this study. First, NJ has 565 municipalities including 254 boroughs, 52 cities, 15 towns, 241 townships, and 3 villages. It ranks seventeenth in the nation for the number of total municipalities (including towns and townships) and thirty-first in the nation for the number of municipalities per capita. Beyond their large number, NJ’s municipalities include great variation in size, structure, and women’s representation, the main variables for this study. In fact, NJ has 466 municipalities with populations under 25,000, cities and towns that would be excluded from many of the national-level studies done on localities to date. Even studies that have used a much smaller cutoff, like Trounstine and Valdini (2008), would miss ninety-five NJ municipalities with populations under 2,500.
NJ municipal councils also vary in partisanship, with 15.8 percent nonpartisan and 84.2 percent partisan councils, and types of election: 500 municipalities (88.5 percent) elect all council members in at-large elections, while the remaining 65 (11.5 percent) elect council members by ward (10) or to a mixture of at-large and ward seats (55) including seven of the top ten most populous municipalities in the state (New Jersey Legislative District Data Book 2014).
Finally, women’s representation varies significantly across NJ’s 565 municipalities. While women represented 22.9 percent of all council members (711 of 3,111) at the time of the data collection in 2014, there was significant variation from local council to local council (mean 22.4 percent, SD = 17.5). While women constituted 50 percent or more of members on 10.3 percent of councils, women’s representation was far scarcer on most councils; 41.1 percent of councils had between 10 percent and 30 percent women in 2014. Almost a quarter (23.9 percent)—or 135—of NJ’s municipalities had no women at all on their councils in 2014 (see also Online Supplemental Figure 1). The municipalities with single-sex councils range in population size from 67 to 92,843 inhabitants; can be found in 19 of 21 counties; include cities, townships, towns, and boroughs; and have a range of at-large, ward, and mixed structures.
These layers of variation within the state in municipal size and characteristics allow for testing multiple potential predictors of women’s levels of local representation from social (demographic or population) characteristics to structural differences. An additional benefit to selecting NJ as a state for in-depth analysis is the investigative precedence offered by Regulska, Fried, and Tiefenbacher (1991), who examined gender differences in NJ’s local political representation a quarter of a century ago.
Finally, choosing a single state in which to analyze predictors of women’s council-level representation allows for some control of certain variables such as statewide culture, reliance on council-level governance and relationships to other levels of government, and the historical representation of women statewide. At the time of the data collection, women were absent from the top elected positions within the state. Although a woman held the office of lieutenant governor, no women served in the NJ congressional delegation. Only one woman, Christine Todd Whitman (R), had ever served as governor in NJ (from 1994 to 2001), and there is no evidence that her election led to increased electoral opportunities for women in NJ. It can be reasonably argued that the dominance of men in statewide and congressional office is pervasive across all of NJ’s counties and localities despite variation in the representation of women at the local level.
While NJ presents distinct characteristics that benefit this analysis, its levels of women’s local representation appear average in comparison to other existing data. No one really knows the level of women’s representation in US local politics (Holman 2015, 23), but van Assendelft (2014) claims that “women comprise approximately 28 percent of city councils” nationwide and Ransford and Thompson (2011) find that women represent 17 percent of council members in New Hampshire and 25 percent of council members in Connecticut. At 22.9 percent, NJ falls well within this range of known representational levels.
Hypotheses: Structural, Sociodemographic, and Institutional Explanations
Variation in women’s representation is often found when countries are compared, and the puzzle as to why women do relatively better in some countries than in others has attracted quite a lot of scholarly interest (e.g., Paxton 1997; Reynolds 1999; Thames and Williams 2013). The finding that variation is also found across local councils, not only within the same country but also within the same state, leads to the no less intriguing question about why this is so. How can the variation in women’s local representation within cases that are quite similar be explained based on statewide characteristics? Some may assume that when comparing local governments within the same state, at least some of the cultural variables would be so constant that the variation in women’s representation would be minimal. However, native New Jerseyans will argue that rural areas in the southern and western parts of the state are completely different from the urban areas separated from New York City only by the Hudson River. Still, to outsiders, these differences do not seem as significant as when, for instance, countries in Scandinavia and the Middle East are compared in analyses of women’s parliamentary representation.
Developing hypotheses about why these differences in women’s representation on local councils in NJ can be observed has relied upon the general women and politics literature as a point of departure. The analysis is specifically informed by three studies which have had local politics as their empirical focus, namely those done by Regulska, Fried, and Tiefenbacher (1991), Trounstine and Valdini (2008), and Smith, Reingold, and Owens (2012). Based on these literatures, three possible explanations are identified which might help to explain the variation in the presence of women across NJ’s local councils.
Structural Variables
First, the structural explanation ties representational outcomes to overall municipal size and degree of urbanization. In most of their analyses, students of local governments focus on municipal size (number of inhabitants) when trying to explain different dimensions of democracy across local units (e.g., Denters et al. 2015). Municipal size has also been included in studies of women’s local political representation, leading to the empirical finding that the larger the municipality, the more women are likely to be elected (Smith, Reingold, and Owens 2012; Trounstine and Valdini 2008). It should be noted, however, that this relationship has been empirically challenged (Antolini 1984, 39; Darcy, Welch, and Clark 1994, 45). It might not be size per se that leads to more women, but rather population density since attitudes toward women officeholders are more positive in the more urban (and often larger) municipalities (e.g., Kjaer and Matland 2014; Darcy, Welch, and Clark 1994, 47). Because of this controversy in the literature, both of these possibilities are addressed in the hypotheses that the larger the municipality in number of inhabitants and the more urbanized the municipality, the more likely women will be represented on the council.
Sociodemographic Variables
Some scholars question whether urbanization itself leads to more women on the councils; perhaps urbanization is primarily a proxy for some other factors. This leads to the second potential explanation, the sociodemographic explanation, which pursues this argument more specifically. Many scholars suggest that women do relatively well at the polls in urban settings not because people live close to each other, but because city dwellers share demographic characteristics that correlate with a more positive attitude toward female candidates such as greater racial diversity, higher educational attainment, and more liberal ideology (e.g., Trounstine and Valdini 2008; Rule 1987; Kenworthy and Malami 1999; Reynolds 1999; Siaroff 2000; Smith, Reingold, and Owens 2012). Relatedly, Palmer and Simon (2008) find that women candidates for Congress fare better in more racially and ethnically diverse districts, leading to the expectation that higher proportions of Black residents (% African Americans) might yield greater support and success for women. Women candidates have also fared better in Democratic districts and states in the United States for nearly three decades, with the partisan gap in women’s representation growing across levels of office (CAWP 2018). This suggests that municipalities with high proportions of Democrats (% registered Democrats) will be more friendly to women. Likewise, Palmer and Simon (2008) find that districts with larger proportions of college graduates are more women-friendly. Based on this existing scholarship, it is hypothesized in this study that women are more likely to be represented in municipalities with higher proportions of black residents, citizens with a college degree, and Democrats.
It is also hypothesized that there will be a positive relationship between women’s council representation and the prominence of female-headed households in a municipality. Regulska, Fried, and Triefenbacher (1991) argue that a higher percentage of female-headed households has a positive effect on women’s election to local office. They note how this measure “captures several aspects of the transformation of American society and the changing position of women” and to “expect that the places with a greater number of women and female-headed households might not only have women who are willing to run for office, but also are more likely to elect a woman candidate” (Regulska, Fried, and Tiefenbacher 1991, 207). There is also evidence that more affluent communities where residents have a higher income tend to elect more women (Welch and Karnig 1979; Trounstine and Valdini 2008; Darcy, Welch, and Clark 1994; Palmer and Simon 2008). Finally, it is hypothesized that the older the electorate—and therefore the more voters who came of age at a time when politics was primarily a man’s game–the less likely that women get elected. More specifically, it is expected that women are less likely to be represented in local offices in municipalities with greater proportions of senior citizens in the electorate.
Institutional Variables
The third and final explanation draws upon many cross-country studies that have demonstrated that political institutional factors also affect women’s descriptive representation. For instance, decades of scholarship indicate that the size of the legislature (here the council size), measured as the number of seats, affects the representation of women since women often do better in larger than in smaller legislatures (e.g., Frederick 2010, 116; Dahl and Tufte 1973, 84; Matland 1998). However, in a recent study of cities with populations over 100,000, Bingle (2016) found no significant relationship between the number of seats at the city council and the level of women’s representation, while another study of county boards actually found a negative correlation between number of seats and percentage of female representatives (Kellogg et al. 2019). These findings reveal that while there is theoretical agreement that larger legislative bodies should yield a greater percentage of elected women, the empirical results have yet to confirm this claim at the local level. What Kellogg et al. (2019) do find, however, is a positive effect of the number of seats on county boards on the election of at least one woman to that board (Kellogg et al. 2019). In other words, increasing the size of local legislative bodies appears to lead to a greater likelihood that there will be some minimal presence of women on boards.
Some scholars have maintained that the size of electoral districts matters more than the size of the legislature; thus, in the case of local politics, it is not the size of the council but the district magnitude that is critical (e.g., Rule 1987, 484; Matland and Brown 1992, 471). Empirical evidence is mixed on this point. Electoral systems may also play a role in creating more or less opportunities for women’s representation. The most recent study of local offices in California shows that women candidates fare better in district-based elections (Crowder-Meyer, Gadarian, and Trounstine 2015), while other research suggests that women fare better in at-large elections (e.g., Trounstine and Valdini 2008). Lastly, partisanship is much more likely to matter in the US context. Women are less likely than men to be self-starters and often need to be asked and encouraged to run such as by party recruiters (Carroll and Sanbonmatsu 2013b). The role of parties in recruiting and/or creating a support infrastructure for candidates may shape whether or not, and where, women run and win. In order to test the influence of parties, the analysis examines whether or not there are any differences in women’s representation across councils with and without partisan elections.
Altogether, then, in this article, it is hypothesized that the more seats on the council, the greater the district magnitude (at-large elections), and the more partisan (i.e., not non-partisan) the election, the more women will be elected. Each of these three possible institutional influences is included in the empirical analyses below.
Data
The general condition that “it is relatively easier to obtain data about nations than about sub-national units” (Lijphart 1975, 167) is definitely true when it comes to studies of women and politics. Data at other political levels are available through easy-to-use databases such as those provided by the Inter-Parliamentary Union on countries around the world (www.ipu.org) and the CAWP on the state- and federal-level representation in the United States (www.cawp.rutgers.edu), whereas this is not the case when it comes to local politics (Kjaer 2019). Calls have been made for a general database on municipalities (Marschall, Shah, and Ruhil 2011), but these calls have not yet been answered. As already stated, this study is therefore confined to one state, NJ, and a database of all 565 local governments in the state. During the summer of 2014, the individual web pages of each municipality were examined to determine the gender composition of each council. 1 Data for the independent variables were obtained from the New Jersey Legislative District Data Book, published by the Center for Government Service at Rutgers University in 2014. 2 More details on each independent variable are provided in Online Supplemental Table 1.
Analysis: Explaining Women’s Descriptive Representation
Explaining Overall Levels of Women’s Representation
In an effort to explain the variation in the percentage of women represented at the NJ local councils, the first model reported in Table 1 includes the two structural, six sociodemographic, and three institutional variables already introduced. 3 Ordinary least squares regression is applied to identify the most influential predictors of women’s overall representation on a council, measured as a percentage of all members. In the second model, which will be commented upon later, logistic regression is applied to identify significant predictors of the simple absence (or existence) of women on a council.
Women on the Council in New Jersey Local Governments Explained by Structural, Demographic, and Institutional Variables.
Source: The Web sites of the 565 municipalities as of July 2014 and New Jersey Legislative District Data Book (Center for Government Service at Rutgers University).
Note: N = 565. OLS and logistic regressions. B (SE). OLS = ordinary least squares.
Levels of significance: *p < .05, **p < .01, ***p < .001.
The first model performs quite poorly, explaining very little of the total observed variance (R2 = .055). Actually, only the percentage of female-headed households is a significant predictor of women’s representation (p < .05); municipalities with a greater proportion of female-headed households are more likely to have a higher percentage of women among their councilors. 4 This analysis indicates that the variation in women’s proportional representation on NJ’s local councils is not explained by the structural, sociodemographic, and institutional variables most often used in explaining differences in women’s representation. 5
This finding is consistent with comparable studies conducted on local councils in the United States (Smith, Reingold, and Owens 2012) and in Scandinavia (Kjaer and Matland 2014), which also return models with modest levels of explanatory power. A very thorough study of women in local politics in Norway and Sweden also provides interesting context for the findings in this article. Wide (2006) finds that the situation has changed substantially in these two countries over times; until the 1990s, there were significant differences between municipalities based on their structural, sociodemographic, and institutional variation, but those relationships have since partially eroded. The explanatory power of the same model has decreased over the past quarter of a century in both Scandinavian countries; in Sweden, the adjusted R2 was .35 in 1979 and .20 in 2002, and in Norway, it decreased from .27 in 1979 to .11 in 2003 (Wide 2006). Wide (2006) proposes that cultural assimilation in geographic localities may explain these changes. She also argues that attitudes toward female politicians are no longer as determined by factors such as the level of urbanization and the occupational makeup of the population as they used to be (Wide 2006).
But how about the United States? Can this trend, where structural and sociodemographic factors are no longer as predictive of the differences in women’s representation as they once were, also be found here? Focusing the analyses on NJ provides a unique opportunity to take a first look into this question because a previous study on women’s municipal representation has been conducted. Regulska, Fried, and Tiefenbacher (1991) tried to explain the variation in women’s representation at the local level using 1988 data from the municipalities in NJ. 6 Their final model, which has the absolute number of women on the council as the dependent variable, achieves an R2 of .247 with the following three statistically significant explanatory variables: female mayor (negative), percentage of female-headed households (positive), and per capita income (positive). When the more recent data presented in this study are used to run a similar model (with the absolute number of councilwomen as the dependent variable and including only these three independent variables), an R2 of .076 is obtained (with the same signs on the effects but with only female-headed households and income significant). 7 Consistent with the Scandinavian study, this comparison of findings in NJ demonstrates how the same model performs less well now than a quarter of a century ago. Although the data used are limited in explaining why the explanatory power of the structural and sociodemographic variables has declined, Wide’s (2006) arguments may prove helpful: the normalization and acceptance of women’s political leadership has increased somewhat over time blurring the municipal-level differences.
Explaining Blanks in Women’s Representation
While these indicators may not help explaining the variation in the percentage (or absolute number) of local council seats held by women, they may contribute to the understanding of why some councils have no women members. As noted above, NJ had 135 municipalities where not a single woman was serving on the council in 2014. These instances of male-only councils can be considered blanks in women’s descriptive representation. Just as it is interesting to know what predicts greater representation of women, it is crucial to understand what explains the presence or absence of women in legislative bodies. Therefore, the analysis is refocused in this section to examine why women are present or absent on local councils.
As noted earlier, the claim is not that having one woman on a governing body is sufficient to ensure that women’s voices are heard in policy discussions, that gender specific problems are infused into the political process, or that the council’s decisions are considered legitimate by female voters. Instead, it is suggested that male-only spaces are the most clearly problematic in achieving these aims for women’s political representation. Thus, seeking any level of representation over no representation is just one step toward political progress for women at the council level. And while councils that had no women in 2014 may have had a woman member at an earlier time, the current absence of women in most cases would seem nevertheless to indicate a governing environment where women’s voices and perspectives are less likely to be heard or valued.
The second model in Table 1 seeks to explain where women are present or absent on NJ’s city councils. Included are the same independent variables as in the first model, but the dependent variable has been changed to indicate whether or not there are any women at all present on the council. Since this dependent variable is dichotomous, logistical regression is applied.
In this second model, explaining the presence of women, the structural variables are again insignificant and the only sociodemographic variable that matters is income. Higher average income among the electorate has a positive effect on the presence of councilwomen (p < .01), supporting the idea that more affluent communities more often elect at least one woman. One institutional variable, the number of seats on the council, also has a significant and positive effect on women’s descriptive representation (p < .001). This finding indicates that the more seats on a council, the more likely it is that a woman will hold at least one of them. Importantly, this finding is very robust; no matter which of the other variables are included or excluded from the model, the number of council seats remains a statistically significant predictor of the presence of one or more women on the council. This second model explains a fair amount of the total variance (R2 = .175).
The finding that the number of seats on a council correlates with having at least one woman serving on council might not come as a big surprise; the more seats, the higher the odds that one of the elected politicians will be from a traditionally underrepresented group. But the fact that the finding is unsurprising does not mean it is unimportant, especially in discussions of possible reforms that might enhance women’s election to political office. Although increasing the number of seats is not commonly included in discussions of structural reforms that might help ensure and expand women’s presence in local politics, the analysis suggests that the effects of this approach could be quite substantial.
Discussion: The Impact of the Size of the Governing Body
When it comes to size, studies of women in local politics—like so many other studies of democratic dimensions of local governments—have been focused on the potential effect of the size of the municipality in terms of number of inhabitants. The study demonstrates that it is not the size of the municipality but rather the size of the council that contributes to whether or not women are present on the council. This possibility has been suggested theoretically and also demonstrated based on simulations in a European context (Kjaer and Elklit 2014). However, this article demonstrates that in the United States, where the number of seats is comparably low, the size of the council really matters. In the Scandinavian countries, for example, women might still be underrepresented on some councils (Kjaer and Matland 2014), but since council size often varies between twenty and thirty seats, councils without any women have been eradicated. Contrary to this, in the United States, the number of seats on the councils is often below ten, and councils with no women are found more often—according to the analyses—in the municipalities with smaller councils.
But how substantial is this effect? Figure 1 presents the predicted probabilities for having at least one woman at the council for all council sizes between three and eleven seats, which is presently the span of seats applied at the NJ local councils (all other variables are kept at their mean).
Predicted probability of having at least one woman on council.
Figure 1 shows that the effect is substantial; if there are three seats on the council, the predicted probability for at least one council member to be a woman is 48.0 percent whereas the probability increases to 98.7 percent for a council with eleven seats. The probability for women’s presence on the council can be almost doubled by increasing the number of seats from three to nine (48.0–95.3 percent). 8
This is indeed a substantial finding, and it is also an important finding since the size of the council is at least somewhat engineerable while other predictors of, or barriers to, women’s representation often are not. Most discussions about the utility of altering institutional arrangements to bolster women’s representation have focused on the implementation of electoral quotas (Dahlerup 2006; Krook 2009), where effectively policed gender quotas are promoted as the ultimate tool for and the ultimate fast track to gender equality (Dahlerup and Freidenvall 2005). But in some countries, such as the United States, the idea of gender quotas contradicts dominant conceptions of merit-based achievement and elections, so much so that quotas are not really feasible tools for changing the gender composition of public officials (Krook 2006). As a result, the future of women’s descriptive representation in the United States boils down to whether other, less radical, electoral engineering tools can be effective or whether more long-term social change will first be necessary. In NJ, the analysis shows that there is an institutional variable to engineer—namely the number of seats on the council—while simultaneously promoting other efforts to support women’s political advancement.
This reform would probably be the most effective means to insure all local governments have at least one female legislator at the council level. By increasing the number of seats on councils, it would be possible to “…open…up new opportunities for women and minorities to serve in the body” (Frederick 2010, 116; see also Dahl and Tufte 1973, 84). However, this solution is not so simple, nor without controversy. First of all, it is not always possible to change the number of council seats immediately, at least not for the municipality itself. Even though NJ is a Home Rule state, there are some limitations on how councils are structured. According to the rules governing municipalities, seven of the eleven possible forms of municipal government in NJ specify how many seats the council should have, though just one restricts this number to one option; the remainder offer multiple options of size, ranging from small to large. 9 For example, in the Council-Manager form, councils are permitted to have five, seven, or nine members. This type of local discretion ensures that increasing the number of seats is possible without forcing a municipality to change their form of government or to apply for a special—self-specified—charter, which is also a possibility in NJ. Secondly, there will be people in some of these municipalities who would be unsympathetic to the suggestion that council sizes should be raised only to increase women’s representation. Increasing the number of council seats might also have disadvantages to the democratic process, such as complicating deliberation and debate, contributing to longer meetings, and/or making it harder for constituents to monitor councilmember behavior. There is a trade-off between representativeness and efficiency (Kjaer and Elklit 2014, 156), and therefore electoral engineers have to “balance the need for adequate representation and legislative functionality” (Frederick 2010, 6) and solve the tension “between representing the voices of many while retaining the ability to govern” (Brooks, Phillips, and Sinitsyn 2011, 1).
Importantly, increasing women’s representation by creating new seats assumes that there will be a ready and willing pool of female candidates to fill them. As noted above, the number of women candidates for office in the United States remains significantly lower than that for men at all levels, contributing to the disparity in representation among officeholders (CAWP 2016). Research on recruitment and selection of women candidates, women’s political ambition, and women’s paths to office has demonstrated the challenges to building and empowering a female candidate pool (Carroll and Sanbonmatsu 2013b; Lawless and Fox 2010). Those challenges might help to explain why the creation of additional state legislative seats through term limits, for example, did not necessarily yield increases in women’s legislative representation (Carroll and Jenkins 2001). Unlike term limits, however, increasing the total number of council seats will not endanger the current women sitting in them and may result in more positive outcomes. A thorough longitudinal study—before and after the increase in council sizes—would be necessary to determine the degree to which these recruitment hurdles would dampen the positive effects of increasing council size.
Regardless, promoting this type of engineerable change will also require parallel efforts to continue to identify, recruit, and support a pool of eligible, interested, and willing women candidates for local office. The findings in the first model also provide an important reminder that increasing the number of seats does not guarantee greater levels of gender parity on councils. Thus, this structural change addresses one aspect of the problem of women’s underrepresentation on local councils, while additional efforts—cultural, structural, and political—are necessary to create conditions for greater gender equality among elected representatives.
Conclusion
Whereas the model seeking to predict the overall percentage of women’s representation on NJ’s 565 municipal councils is limited in its capacity to explain the observed variation, the above analyses have demonstrated that the 135 “blanks” in local governments are most likely to occur where the number of seats on the council is low. In NJ, the size of the councils varies between three and eleven seats. The probability for finding a woman in at least one of the seats increases greatly when moving from the smallest to the largest councils (controlling for a number of structural, sociodemographic, and other institutional variables). This finding might seem obvious to some but is nevertheless significant due to its prescriptive possibility; the number of seats on municipal councils is somewhat engineerable, unlike other variables deemed significant in predicting women’s representation. Local councils can often increase their number of seats, thereby raising the chances that a woman will be elected and that some of the aforementioned blanks will be filled.
This article has identified, at least in the case of NJ, not only a problem (the blanks for women) but also a possible solution (increasing the number of council seats). Local governments in NJ should not be expected immediately to raise the number of seats on their councils to increase the likelihood that women are elected. While they may recognize this action as solving one problem, expanding council size might also create new problems such as lower legislative efficiency. Again, then, this poses a more philosophical and motivational question for those with the power to engineer change: how will they balance the need for women’s descriptive representation with other considerations around representative democracy? Another problem is that creating political opportunities by adding council seats will not ensure women’s representation if barriers to women’s candidacy to fill those seats persists. Promoting greater inclusion at the recruitment and nomination stages remains crucial to increasing women’s political representation (Kjaer and Kosiara-Pedersen 2019).
The study reported here should also encourage other scholars to “go local” in their investigations of women’s political representation, either by conducting similar analyses in different states or by seeking to identify other potential predictors of women’s presence on councils that are not available or irrelevant in NJ. For example, as some municipalities look to change their voting methods to ranked-choice voting, there may be greater opportunity for evaluating variance in election methods across municipalities as a potential influence on the number of women elected to local office.
Most certainly, increasing the number of local council seats will not completely solve the problem of women’s political underrepresentation. Nevertheless, the persistent dearth of women in elected office requires a multifaceted approach that capitalizes on any and all opportunities for change, and as part of a multifaceted approach, engineering greater numbers of opportunities for women to run for office at the local level would address one important constraint on women’s representation.
Supplemental Material
Supplemental Material, SLGR_18-0064R2,_Figure_1_Supplement - Council Size Matters: Filling Blanks in Women’s Municipal Representation in New Jersey
Supplemental Material, SLGR_18-0064R2,_Figure_1_Supplement for Council Size Matters: Filling Blanks in Women’s Municipal Representation in New Jersey by Ulrik Kjaer, Kelly Dittmar, and Susan J. Carroll in State and Local Government Review
Supplemental Material
Supplemental Material, SLGR_18-0064R2,_Table_1_Supplement - Council Size Matters: Filling Blanks in Women’s Municipal Representation in New Jersey
Supplemental Material, SLGR_18-0064R2,_Table_1_Supplement for Council Size Matters: Filling Blanks in Women’s Municipal Representation in New Jersey by Ulrik Kjaer, Kelly Dittmar, and Susan J. Carroll 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.
Notes
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.
