Abstract
Representation is a hallmark of democratic governance. Widely studied within traditional modes of governance, representation is less studied in alternative governance settings, such as collaborative governance arrangements. Collaborative governance arrangements are specifically designed to encourage inclusion and participation among a diverse array of stakeholders in some part of the policy process. Our research contrasts different forms of representation observed in a collaborative governance arrangement and identifies factors contributing to observed patterns in representation therein. We analyze descriptive representation (i.e., “representation in form”) or substantive representation (i.e., “representation in practice”) and look for inconsistencies between them. Our case study is a regional food policy council located in the Western United States. Among our findings is that discrepancies between descriptive and substantive representation can be explained by shared goals, local norms, organizational structure, and heterogeneity in member capacity. We conclude our article with a discussion of the theoretical and practical implications of this research.
Introduction
Representation is a critical component of governance in democratic societies (Hero & Tolbert, 1995). Fundamentally, representation refers to the extent to which constituent characteristics are reflected in governance structures, processes, and outputs (Gazley, Chang, & Bingham, 2010). Scholars have devoted considerable attention to understanding different forms of representation and their implications for traditional modes of governance (Bratton & Ray, 2002). However, the concept of representation has been understudied in settings where governance is often specifically designed to encourage broad participation, for example, in the context of collaborative governance. In collaborative governance settings, diverse sets of stakeholders are engaged in public policy making and management to encourage inclusion and participation in the policy process along with harnessing diverse stakeholders’ various expertise, resources, and support (Emerson & Nabatchi, 2015a).
In studies of collaboration, diversity in representation is often accounted for in terms of the variety of stakeholders formally granted access to a collaborative process. Formal membership is a useful and valid way of assessing representation in a descriptive sense, that is, for describing which of the various stakeholders who have a vested interest in the topic or the work of a collaborative are represented therein. Formal membership is, however, not useful for understanding representation in a substantive sense, or the extent to which those stakeholders included in a collaborative are represented in its actual process, goal setting, and outputs. Understanding which actors are substantively represented in a collaborative process is as important as understanding who can officially participate in it.
We identify asymmetries between descriptive and substantive representation and factors that inform this asymmetry in a collaborative governance arrangement case study. We operationalize descriptive representation as group membership. We operationalize substantive representation in three ways: (a) attendance in a collaborative process, (b) participation in a collaborative process, and (c) agenda issue presence, or the extent to which the issues that are the foci of members are reflected in its formal agenda. The contrast between descriptive and substantive representation can be thought of in terms of the distinction between “representation on paper” and “representation in practice.” Discrepancies between descriptive and substantive representation are particularly noteworthy in the context of collaborative governance arrangements because these arrangements are often designed to encourage balanced representation of diverse stakeholders (Ansell & Gash, 2008; Connick & Innes, 2003). Discrepancies between descriptive and substantive representation prompt questions about which factors specific to the collaborative process or to members contributes to discrepancies.
While collaborative governance scholars assert that representation is important to collaborative governance structures, there is no scholarship that specifically investigates descriptive and substantive representation in juxtaposition. In response, we address the following three questions in this article: (a) How can descriptive and substantive representation be measured in the context of collaborative governance? (b) To what extent is there variation between descriptive and substantive representation in a collaborative governance arrangement? and (c) What factors contribute to differential representation among diverse types of actors?
Our research setting is a food policy council (FPC) located in the Western region of the United States, referred to hereon as the “WFPC.” FPCs are examples of a collaboration increasingly seen in food system governance. FPCs are comprised of representatives from different food system sectors that come together to engage in food system planning (Siddiki, Carboni, Koski, & Sadiq, 2015). The WFPC is a well-established, public entity with dedicated governmental support but is structured to include representatives or participants from two dozen distinct local food system stakeholder groups including public, private, and nonprofit groups. Within the context of an FPC, like the WFPC, the concept of representativeness is measured as the extent to which participants designated to reflect the interests of different food system stakeholder groups are represented practically and substantively in the process and/or outputs of the group. Like most other FPCs, the WFPC not only makes explicit efforts to include representatives from groups affected by food system issues but also struggles with discrepancies between descriptive and substantive representation. The WFPC confronts issues common to other FPCs and collaborative governance arrangements.
Data for our WFPC case study come from a yearlong coding of council rosters and meeting minutes buttressed by in-depth interviews with council participants. Coding council rosters and meeting minutes allows us to compare descriptive representation with substantive representation, while the interview data help to explain variations in participation within the council, and beyond council design elements. Our study contributes to a better understanding of the complex and fluid landscape of food governance in the United States, and underscores the challenges related to the disconnect between design and implementation of collaborative governance arrangements.
Collaborative Governance and Representation
In recent decades, collaborative governance has emerged as a popular alternative to traditional modes of governance. In collaborative governance, diverse arrays of stakeholders from the public, private, and nonprofit sectors are convened for one or more public purposes, including policy making, policy implementation, or coordinating public service delivery tasks (Emerson, Nabatchi, & Balogh, 2012). Collaborative governance is deemed an effective strategy for dealing with wicked problems, or problems without readily available solutions. Proponents of collaborative governance argue the complexity and uncertainty that characterize wicked problems are best addressed through engaging those most directly interested in and affected by them, including individuals with relevant expertise (Emerson & Nabatchi, 2015a).
In a handful of studies, scholars have explicitly sought to understand the role of representation in collaboration, verifying its importance in enhancing the perceived legitimacy of collaborative processes (Beierle & Konisky, 2001; Emerson & Nabatchi, 2015a; Weber, 1998) as well as in fostering outcomes like participant learning (Muro & Jeffrey, 2008). Leach (2006) describes a representative collaborative process generally as one that “ensures that the interests of all affected individuals are effectively advocated, either in person or through proxies” (p. 101). Even though representativeness is assumed to be a salient component of collaborative governance (Ansell & Gash, 2008; Connick & Innes, 2003; Emerson & Nabatchi, 2015a), many studies treat it as an assumed feature and antecedent of collaborative outcomes. Little attention has been given to how to operationalize representation as a multidimensional concept to reflect the different forms representation takes in collaborative fora. Relatedly, questions about the implications of different forms of representations have yet to be fully investigated.
Further insight on representation in the context of collaboration and formally organized venues may be drawn from scholarship on organizations and political institutions. In considering the link between nonprofit board diversity and external collaboration, Gazley et al. (2010) conceptualize representation as a “purposeful term, referring to the extent to which an organization reflects constituent characteristics in its governance or operations” (p. 610). Political organization scholars have further delineated the concept of representation into (a) descriptive or passive representation and (b) active or substantive representation (Keiser, Wilkins, Meier, & Holland, 2002; Pitkin, 1967). 1
In the context of research on political organizations, Pitkin (1967) delineated descriptive and substantive representation by describing the former as mirroring social demographics and the latter as reflecting the interests of those being represented (Swers, 2005). Drawing a similar distinction, Keiser et al. (2002) describe passive representation in terms of how demographically representative a group is relative to the larger population. Active representation is the extent to which the interests of groups within the larger population are represented in public processes. Active representation is also similarly defined in terms of whether the interests of the demographic groups represented within a venue (e.g., bureaucracy, congress, etc.) by individuals with like characteristics are reflected in the outputs of that venue (Wilkins & Keiser, 2006). Wängnerud (2009) explains that descriptive representation is assessed based on the number of women serving in national parliaments, whereas substantive representation focuses on the impacts of their representation therein.
Beyond describing and contrasting descriptive and substantive representation, a few scholars have investigated which factors enable or inhibit substantive representation. Much of this work is directed at testing a common assumption that descriptive representation implies substantive representation; that is, so long as certain stakeholders are participating in a governance venue, their interests will be represented in the outputs of the venue. Examples of factors posited to influence substantive representation include the following: social identity, institutional and political opportunity structures and contexts (Reingold, 2006), stakeholders’ role or positions within a governance arrangement (Swers, 2005), and the number of stakeholders from a particular group represented (Beckwith, 2007). Wilkins and Keiser (2006) also emphasize that discretion to meaningfully participate in a venue or process is a necessary precondition for the translation of passive to active representation.
Referring specifically to factors that may hinder representation in collaborative contexts, though not speaking directly to the link between descriptive and substantive forms of representation, Leach (2006; see also Ansell & Gash, 2008) points to the salience of resource constraints. Leach highlights the observed reality that among participants in collaborative processes, there tends to be notable variation in the time they can devote to participation, distances they must travel to participate, remuneration for participation, and even level of knowledge of issues germane to the collaborative process (Leach, 2006). Generally, resource constraints, particularly when there is an imbalance in such among participants, are expected to compromise overall representativeness (Mostert et al., 2007). In addition to resources, several scholars of collaborative governance have pointed to the role of leadership in maintaining balanced representation in collaborative processes (Ozawa, 1993). According to Ansell and Gash (2008), leadership can be particularly important when power and resources are “asymmetrically distributed.”
Other factors that may influence active participation 2 in collaborative settings are highlighted by Koehler and Koontz (2008) through their study of collaborative watershed partnerships. Koehler and Koontz are specifically interested in exploring what contributes to individuals actively participating in partnerships once they are already members. They find that gender, occupational affiliation, open communication, adequate technical support, and agreement among participants all matter (Koehler & Koontz, 2008). Koehler and Koontz’s study provides a useful foundation for further investigation into the differences between inclusivity in a collaborative process and substantial representativeness therein.
The extant scholarship on representation informs terminology used in this article. We define descriptive representation, in the context of collaborative governance, as formal representation of the various stakeholders that have a vested interested in the topic or work of a collaborative. We refer to substantive representation as representation of these stakeholders in the process, goals, and outputs of the collaborative process. We conceptualize substantive representation along two broad dimensions: procedural or process representation and output or outcome representation. Procedural substantive representation can manifest as attendance in a collaborative process and participation in a collaborative process. Output or outcome substantive representation can manifest in agenda issue presence, or the extent to which the issues that are the foci of collaborative members are reflected in its formal agenda.
Setting and Case: Food Governance
Food system governance challenges are typical of other collaborative arrangements: exceedingly complex; fraught with global, national, and local issues of growth, distribution, equity; requiring balanced consideration of interdependent social, economic, and environmental issues, and so forth. Attempts to grapple with systemic food issues have largely occurred at the local level in the United States, where a diversity of stakeholders from a variety of sectors come together to address issues relating to agriculture, nutrition, social justice, commerce, and environmental protection. Top-down efforts to understand food systems are virtually nonexistent, and resources are generally scarce, yet demand for systemic thinking about food-related issues is growing. Collaborative governance is emerging as the dominant choice for food systems. Food systems are thus excellent contexts for studies on representation in collaborative governance.
An emerging governance structure that performs a variety of tasks related to the local food system is the FPC. Although there is no universal council model, most bodies contain representatives from different areas of the local food system. Representatives are typically convened to provide advice on existing and emerging systemwide issues, and occasionally, policy development. In this study, we collect data from one FPC over 1 year.
To protect the anonymity of our case, we refer to the FPC we studied for this research as the Western food policy council, or WFPC. Formally, the WFPC can be classified as a committee within a metropolitan planning organization. The WFPC, founded in 2010, began as a local FPC that served one county between 2006 and 2009. This local council approached a regional governance and planning body when members became interested in expanding the council’s service area. The regional council provided the seed money for the FPC after recognizing that food policy fits well with the work of the regional council. To ensure a smooth transition from the FPC to the WFPC, the regional council created a transition committee that developed the structure of the WFPC. The transition committee also gave the WFPC its current name and came up with 20 seats (initially), to include two cochairs (one has to be elected), a vice chair, ordinary members, and some alternates. The WFPC is a committee nested within the regional council. The WFPC has a steering committee, and both the WFPC and the steering committee meet once a month (the steering committee is not under study). The WFPC now offers its services to the four-county region, with a combined population of approximately 4 million as of July 1, 2015.
The mission of the WFPC is to develop just and integrated policy to enhance public health and regional food systems. In addition, the WFPC serves as a platform for engaging a diverse membership from businesses, local jurisdictions, institutions, and community organizations. Some of the challenges facing the WFPC are food access, preservation of agricultural resources, and the resilience of the food system during disasters. Much of the funding for the WFPC comes from external sources—the city, county, and state governments—through project-based contracts. The WFPC’s role in the policy process is to serve as a linkage between like-minded organizations, to generate information regarding the food system for policy makers, and to identify issues on the food policy agenda of the region. Since its inception, the WFPC has completed several projects, including a report assessing the viability of farmers’ markets, a food policy blueprint, comprehensive planning for urban agriculture, and the establishment of a database of food-related policies within the region. The WFPC is an appropriate case to investigate representation in collaborative governance for several reasons. First, the WFPC consists of a diverse group of food policy actors such as elected officials, tribal leaders, regional public health and economic development officials, food industry representatives, and a variety of public and nonprofit organizations. Second, the WFPC is located in a region known for collaborative policy making and innovation in food policy. Third, the WFPC maintains a publicly accessible repository of council documents amenable to coding for our analysis of council representation. Finally, there is precedent in using single case studies of multicounty FPCs (Johnston, Nicholas, & Parzen, 2013) in investigating governance questions.
Data Collection
Data Sources
Data for this research were collected to identify descriptive and substantive representation in the WFPC and factors that influence members’ substantive representation. Our study period covered council meetings from July 2013 to June 2014 and followed with interviews in August 2014.
To determine descriptive representation, we identified actors formally granted access to the WFPC through membership. We created a member list largely from individuals found in a significant 3-year planning document published in October 2014 (2 months after our interviews). This action plan is featured prominently on the council’s website and, importantly, is the only public document that contains a list of members (other than individual meetings). The action plan was a collective document for the council and sets the stage for new initiatives. 3 Twenty-four members were identified from this data collection process.
To determine substantive representation, we collected data from WFPC meeting minutes and conducted hour-long semistructured interviews with council participants. We reviewed publicly available meeting minutes from the WFPC for 10 meetings (all that were held) spanning the study period. During our study period, the same staff member took notes at each meeting. The organization meets 10 times per year for about 2 hr in a centrally located facility. Because the organization spans multiple jurisdictions in a major metropolitan area, some members will occasionally call in. However, the vast majority of participation occurs in-person at meetings that are open to the public. The meeting agendas are published in advance. Minutes record discussed topics, attendees (members and nonmembers), and participants in discussions.
Interviews were conducted with a sample of WFPC members in August of 2014. An invitation to participate in an interview was sent to all council members after the project was introduced to members by the council coordinator following communication with one of the authors. Thirteen of the 24 members of WFPC volunteered and were interviewed. In the interviews, we asked respondents to characterize a number of issues related to FPCs broadly and the history of the WFPC. Specific to this research, we asked respondents questions regarding the influence of organizational structure on process and representation of different stakeholders in the WFPC. The interviews were designed to provide a reliability check on the use of minutes and to assist in helping to explain variation in representation and participation in council activities.
Data Collection: Meeting Minutes Coding
Meeting minutes provide the foundation of our substantive representation data. Meeting minutes were coded specifically to collect information on organizational attendance and participation in meetings to capture substantive representation in process, while our coding of agenda items captures substantive representation in outputs.
Substantive representation in process: Organizational attendance and participation
In determining attendance, we looked at which council members were actually present (in-person or by phone) at each of the 10 meetings. The focus of our work is the role of organizations as representative of interests; thus, we measure organizational attendance rather than individual attendance. We choose organizational representation over individual representation because of FPC characteristics. In the WFPC, there are a number of elected officials or high-ranking administrators who are intended to represent various jurisdictions of government. In many of these cases, we found that the actual members rarely attend; however, the organization is structured to contain alternates who, in many cases, are the de facto representatives of member organizations. Choosing organizations enables our research to absorb changes in individual representatives (which occurs occasionally) of organizations as well as attendance by alternates. As the FPC membership is intentionally structured to represent particular interests and types of organizations, changes to individual members do not result in significant membership changes for the council. The council replaces departing members with new members of similar type and focus (this occurs rarely during the study period).
We consider an organization as attending a meeting if any representative of that organization attends a meeting. We use a simple binary code for meeting attendance: either an organization attends a meeting or it does not. We use the same binary coding scheme for meeting participation: either an organization makes a statement for a unique agenda item or it does not. Thus, organizations can attend meetings but not participate; however, the converse is not possible. We do not make a distinction for participation extent by agenda item. Table 1 identifies the types of organizations represented in the WFPC as well as the foci of the organizations represented in the council.
Coding criteria for Organizations Participating in the Food Policy Council.
We coded organizational participants of the WFPC by type and area of focus. We used nine criteria to describe organization types: nonprofit sectors, special districts (in this case, conservation districts and other special authorities), tribal members, universities, private businesses, three jurisdictions of government (city, county, and region), and a special category for chairs (who are “at large” members). 4
Substantive representation in output: Issue representation
We coded meeting minutes to capture substantive representation in outputs by looking at variation in the agenda of each meeting. We group minutes from each meeting into distinct topic statements. As an example, “Council members suggested follow-up that advances food-related economic development with regional economic development entities. Additional research documenting the dollars generated by the local food economy would be useful.” This statement is coded as a distinct topic, related to economic development. The statement is from a local government representative; thus, we attribute the statement to the member (type—city, focus—local department) who is guiding this discussion (for the process measure mentioned in section “Substantive representation in process: Organizational attendance and participation.”). Table 2 identifies the topical categories in accordance with which meeting minutes were coded. We coded 139 such statements over 10 meetings using 16 topic codes that have been used in previous FPC-related research (e.g., Siddiki et al., 2015). 5 There are 39 unique participants in meeting minutes, including nonmembers. Coding nonmember participation helps to separate descriptive and substantive representation from agenda representation.
Topic Coding Categories for Statements During Meeting Minutes.
In addition to topic, we coded each statement based on the extent to which the information dominates the meeting agenda, what we term significant. Much of the content of the council is in the form of an announcement (e.g., upcoming events such as a farm tour). Other types of content offered by council members include updates from representatives of programs within or outside the council. We coded significant statements as those that comprise extended discussion of substantive issues (not announcements or program updates) involving other members of the council. Of the 139 total statements, we characterize 33 as occupying a “significant” portion of the agenda.
Data Analysis
We use a Shannon diversity analysis to compare diversity by type of representation in organizational design, meeting attendance, and meeting participation. We also use a Shannon equitability analysis to assess the extent to which participation is concentrated or dispersed across members at meetings. Qualitatively, we identify common themes from interview responses to issues related to FPC organization and specifically concerning representation of members.
Diversity and Equitability Analysis
We assess the difference between descriptive and substantive representation using an entropy score. In this case, we choose to use a Shannon–Weiner index (or Shannon’s H), common to characterizing species diversity in ecology. Shannon’s H has a range of zero to the natural log of the number of populated categories in the data set, represented here:
where the proportion of cases (i) relative to the total number of cases (pi) is multiplied by natural logarithm of this proportion and summed. The index gives a general sense of the diversity of elements in each category (here, organization type, organization foci, and discussion topic).
To further investigate the relative equality of representation versus the quality of representation, we conduct an equitability analysis of the data using Shannon’s EH. Shannon’s EH is similar to other entropy statistics (e.g., Blau scores or Herfindahl indexes) in that the statistic represents a transformed proportion between 0 and 1. Shannon’s EH is the H score transformed as a function of the number of populated categories in each variable. Mathematically, the expression is given as follows:
This is an important distinction as in many of our cases, there are observations of 0 (e.g., private actors making significant comments or representatives of several of the organizational foci). Thus, some of the potential statistical anomalies observed in the comparison of H scores are smoothed in a comparison that is less sensitive to the number of cases in each variable (Shannon’s EH)—a key challenge when assessing diversity as a function of categories associated with different variables, the variance of which is in part due to the authors’ coding choices.
Results
Who Attends?
Attendance is a basic form of representation in council meetings but a crucial one given that participation in a collaborative governance organization ostensibly occurs by members sitting at the table. Hence, attendance is critical in understanding descriptive versus substantive representation. Our coding scheme enabled us to analyze attendance as a function of organization type and organization foci. Figure 1 displays the rate of attendance by organization type (1a) of those actually listed (1b) by the FPC. This does not include other individuals or organizations who attend meetings. The pie chart on the left (1a) is the actual construction of the 24-member council; the pie chart on the right (1b) represents the actual attendance of stakeholders in each group (each attendee gets a count of “1” per meeting; 148 attendance observations over 10 meetings out of a possible 240 attendance opportunities). 6
Member attendance.
Differences that are observed in the two previous charts are enhanced in examining the average rate of attendance. An initial observation from these data is that the council is designed to be a roughly equal distribution of nonprofit (29%) and governmental (30%) sectors, but the reality of attendance is that city and county officials make up a greater share of all meetings (22% for nonprofits and 36% for government members), and have a much stronger average rate of attendance.
Next, we look at council representation by foci, shown in Figure 2. The pie chart on the left (Figure 2a) examines the representation of member foci in the design of the FPC—descriptive representation. The council tilts toward local departmental descriptive representation by design; however, aside from this, there is a broad distribution of interests. The chart on the right (Figure 2b), however, shows the representation of these interests in practice as a function of attendance—substantive representation. Observations in this chart are counts of organization attendance in meetings by foci. In practice, governmental interests have stronger substantive representation (36% of total) than descriptive representation (29% of total). 7
WFPC member representation by foci.
Identifying Variation in Council Representation Diversity
Our research intends to uncover the relationship between descriptive representation and substantive representation. Figure 3 presents a comparison of representation in the design of the WFPC (descriptive representation), member and nonmember participation (substantive representation in procedure), and the content of the WFPC (substantive representation in outputs).
Heterogeneity in FPC representation.
The “roster” column corresponds with descriptive representation (information contained in Figures 1a and 2a). We measure distinct components of substantive representation in the remaining columns of Figure 3. The “attendance” column is the barest conceptualization of substantive representation in procedure, corresponding with the presence of member organizations in meetings (information contained in Figures 1b and 2b). The “participate” column is conceptualized as a stronger version of substantive representation in procedure, measuring organizations’ participation in WFPC meeting minutes. The “significant contributions” column is conceptualized as the strongest version of substantive representation in procedure, measuring the contribution of member organizations to items that receive substantial council discussion.
Across each of these columns, we use the two measures of organizations we conceptualize in Table 1, organization type (solid line) and organization focus (dashed line). Substantive representation in outputs is contained in the dotted line marked “topic.” The topic line corresponds with heterogeneity of topics raised in meetings from participants (see Table 2 for this list). Figure 3 presents a combination of these data using Shannon’s H to create common diversity scores for descriptive (organizational design) and substantive representation (organization presence and outputs) plotted against variations in diversity across organization type and foci. The Shannon’s H score is best for comparisons within each of the organization type, organization focus, and topical agenda items; for our purposes, it is inappropriate to compare scores across these functions (though the trends are useful). While there is much to unpack from Figure 3, two trends are clear: (a) substantive representation is less heterogeneous than descriptive and (b) significant outputs are from less heterogeneous sources than nonsignificant outputs.
Identifying Variation in Council Representation Diversity
While the diversity analysis can compare changes within organization type, organization foci, and topic, direct cross-categorical comparisons are hampered by the different bases of each type. Nevertheless, we use Shannon’s EH as an entropy statistic which converts the scores to a common scale (see Figure 4). The equitability scores additionally contrast with the diversity scores because they measure equality of contribution across actual contributors (equitability) rather than potential contributors (diversity). The number of contributors and contributor types dwindle when the data set is shrunk to look only either at members or at significant participation. Comparing the two statistics, H (Figure 3) consistently drops as representation becomes more substantial, which is what we might expect; however, EH (Figure 4) dips from roster to attendance to participation but rises for significant contributions.
Representation equitability in the WFPC.
The “dip and rise” represents a key point for our explanation. While Figure 3 shows that the overall diversity of members shrinks as representation becomes more substantive, Figure 4 indicates that the equitability score for significant contributions is much higher than for participation—meaning that there are many organizations, who generally participate, but a few members dominate the overall discussion; however, within the group that offers significant contributions, there is greater balance across contributors. Put another way, the results in Figure 4 suggest a kind of “council within a council” which equally shares the load as a small group. This small group of actors’ influence is enhanced by consistently high attendance at meetings in contrast to other organizations on the council; however, high attendance does not guarantee high levels of participation as there are organizations that attend with greater frequency but participate less in meetings.
Factors Influencing Substantive Representation: Interview Data
We employ a mixed-method analytical strategy for understanding representation in the WFPC. We relied on descriptive, quantitative analyses of data obtained through a coding of council documents to identify how different council members were descriptively and substantively represented in the Council. Our interpretation of these descriptive, quantitative analyses focused particularly on identifying discrepancies between descriptive and substantive representation. We relied on interview data for insight on factors that influence any observed discrepancies revealed through our descriptive, quantitative analysis. Our analysis of interview data involved determining common trends, or themes, across interviewees’ responses. Based on this descriptive, qualitative analysis of interview data, four themes emerged that help elucidate the differences we observed in descriptive and substantive representation using the quantitative analyses described above. Taken together, the quantitative analyses helped us understand trends in representation in the WFPC and the qualitative analysis helped us understand what factors inform these trends.
Our data characterize the WFPC as a collaborative arrangement that is structured to represent a diverse array of interests in the local food system, yet one that displays less substantive diversity in actual participation in governance. The meeting minutes offer little explanation on potential causes for group attendance or participation. We turn to interview data of 13 members of the WFPC to uncover dynamics that contribute to representation in the WFPC. Members were asked to identify factors that contributed to collaboration within the organization broadly related to goal agreement among participants, openness of the organization to a diversity of topics, and organizational procedure related to agenda-setting and participation. Interview responses were free-form and wide ranging regarding the organization. We conducted an inductive analysis of these responses. Four dominant themes that shape the participation environment in the WFPC emerge from our inductive analysis of these interviews: shared goals, local norms, organizational structure, and heterogeneity in member capacity.
Shared goals
Representation in a collaborative governance arrangement is in part a function of members’ desires to participate but also in the openness of the organization to hear their contributions. Each of these issues emerges in interviews with participants as influential in determining participant contribution to food system governance.
In general, we begin with the expectation that a greater proportion of shared goals will contribute to greater ability and desire to contribute. “Shared goals” can be thought of as members of the council bringing similar goals to the table and/or agreeing to similar collective goals for the council. On the surface, the council would seem to exhibit strong goal sharing across members and for the group itself (e.g., improvement of the food system). However, this is generally not the case. Given that the council is designed to be diverse in terms of which aspects of the food system it intends to address, we would expect that people coming to the council would have a very diverse set of goals. Thus, members do not join with the same level of expertise and shared experience. The onus to encourage participation and to create a shared set of goals so that participants have a common sense of purpose falls to the council. This means that the council needs to show that it has a clear mission, that people understand this mission, and that the council is the primary vehicle for achieving this mission.
Interview data show a range of statements regarding goals and goal clarity. Each member envisages multiple priorities for the food system—some overlapping and others independent. However, uncertainty regarding the overall purpose of the council is widespread. A local governmental member noted, “But in terms of what the council actually does, I’ve sat there trying to figure that out, to be honest” (Interviewee 4). Another interviewee indicated a catch-as-catch-can mission that is a function of disparate individual preferences rather than a coherent strategy:
We’re not strategic. We’re all over the board. We’re stuck with that we want to do everything. But we can’t. And where do we start? And people, I think, get slighted if their issue’s not in the [mission of the organization]. (Interviewee 6)
When asked about the role of the council in addressing food issues, one respondent simply stated, “I’m still figuring that out” (Interviewee 13). The following exchange perhaps best sums up members’ uncertainty regarding organizational goals and strategy:
So what do you see the role of the council being in the context of the issues that you’ve just discussed?
Well I think it remains to be seen to a certain extent. I think there’s a very strong orientation early on to just explore the issues, introduce the players, foster communication. And I’ve talked about it as this spontaneous combustion theory, that if you put the proper ingredients together, suddenly the fire will emerge. And I don’t think there is any thought, not much thought given to how to fan the flames. And I think we’re in an interesting transition period with the [sic] WFPC. To determine whether it’s going to become a more activist organization. And there’s some sentiment among policy council members that it should go that way. And other people feel that it really shouldn’t; that it should continue to be a forum for conversation and education. So it remains to be seen, what’s going to come of all this.
At the same time, it must be noted that the purpose of some members on the council is to learn about issues that emerge in the local community. All interviewees valued, to some extent, the importance of open-ended discourse regarding food issues, which are understandably diverse. One member specifically noted that without the council, he or she would not know what is going on: “I do think they do a good job of even broadening just a kind of—knowledge of the group. Which I think is important (Interviewee 11).” Interviewee 2 also noted, “Really diverse group of people who are really open-minded and that feel comfortable speaking about the issues they are comfortable with . . . everybody is totally willing and open to educate themselves.”
Local norms
Local cultural norms regarding process are pervasive through the interview data, suggesting that part of the council’s success in attendance is a function of an emphasis on hearing diverse perspectives. Such emphasis on incorporating diverse perspectives is corroborated by Interviewee 2: “The council often invites people in the community to present on some issues to figure out how people can collaborate and get to know the work of others.” In general, interviewees laud the council’s sensitivities, but there are some that suggest that the council is held back from addressing substantive issues. “Right now, we’re really conflict averse” (Interviewee 1). Interviewee 4 also notes that the process focuses more on consensus building through niceties than conflict: “Here the process is that people are nice to each other and sometimes it drives me absolutely batty. Because it takes a long time to get anything done.” One member who spent time in a previous governmental position strongly describes the process: “I’ve been startled by how much stock we put into really elaborate, thoroughgoing stakeholder processes to try to get to a result . . . it’s unprecedented for me, how much time we spend doing it” (Interviewee 8).
Few entrepreneurs emerge to guide the council beyond what many see as informative, if benign, discussion. One interviewee states: “You need a champion” (Interviewee 3). Another interviewee agrees that “sometimes I think you need a little bit of the squeaky wheel on the council” (Interviewee 5). The longtime chair is viewed as the principal entrepreneur of the council by nearly all interviewees, and the representatives of the regional government body in which the WFPC is located are viewed as key organizational cogs. Regarding representation, then, members attend meetings because of norms of fairness and fairness-favoring candidate recruitment. However, we see substantive participation distributed across a narrow, vocal range of members who include WFPC founders and regional governmental representatives who have particular incentives to ensure organizational function (regional government representatives are not formal members of the council). This is evident in the outsized (in raw and percentage terms) representation of regional governments and the chair in meeting presence found in Figure 5, thus corroborating equitability findings in Figure 4.
Organizational participation in WFPC agenda by organization type.
This conciliatory nature helps explain a substantial council focus on “council governance,” the majority of which are related to goals and future recruitment of the group. As the process of the council supplants the substance of the council, we note some frustration among council members who would like to see the WFPC act more aggressively. Our coding of agendas confirms the council focus from substantive issues to process issues over time.
Organizational structure
Interviews suggest that the WFPC is populated with members who get along, but a key element to representation is the extent to which the council is open to hearing all the food-related issues in the region it serves. Representation for such a diverse range of groups would be difficult on a council favoring a narrow range of topics. We asked interviewees a series of questions regarding the openness of the council to other perspectives.
In general, the consensus is that the council is open (Interviewees 1, 2, 6, 7, 8, and 11) to other groups, while there is some healthy disagreement (Interviewee 10). This openness leads to the lack of organization focus we identify in previous discussion, but for some members, it is the precise reason for the council’s existence—as an information clearinghouse of food-related issues for the region (Interviewees 3 and 11). One member worried that “equity voices [are] not strong—when they show up they are very loud. Their access issue is a problem. They are not underrepresented, but they just can’t make it [to] the meeting” (Interviewee 9). Another member noted that the culture and structure of the organization are open, but the fact that the WFPC had reverted to conducting most of its substantive discussion in the whole meeting (rather than subcommittees) leads to less actual openness: “It is hard to hear [members] equally” on “larger issues without involving many people like 20 or so” (Interviewee 10). Nearly all interviewees expressed some concern that the council does not or simply cannot hear all of the issues facing the region’s food system, though this is largely not thought of as the fault of the WFPC. This structural openness helps to explain the findings that a range of issues are represented in the council, even when a range of individuals might not be (“Topic” findings in Figures 4 and 5).
Heterogeneity in member capacity
Intended versus actual representation is in fact a function of the capacity of members to attend meetings and consider issues before the council. Meeting attendance is crucial given a number of statements by members that collaboration does not occur much outside the council—at least for council-specific issues. Thus, it is imperative that members attend to be represented in council governance; however, not all members have the same organizational capacity to attend.
Nonprofits and special districts, particularly underrepresented in meeting participation in comparison with their roster presence, have smaller staff and fewer resources. Private businesses have a good record of attendance reflecting (in this case) a larger organizational presence but are less influential in dictating the organizational agenda of the WFPC. Governments not only have the capacity to send representatives but also have the capacity to send alternates. Thus, capacity presents a significant hurdle to representation in the council. Interviewee 7 presents the classic trade-off argument succinctly:
A major issue in the region is that there is a lot of competition for resources; in particular, there are a lot of nonprofits. [Our group] tried to get grants through the Council to do more assessment. Participation can be difficult because there is a cost to participation and you don’t know what will come out of the process.
Similarly, Interviewee 2’s statement captures the resource constraint issue aptly: “Some council members from smaller offices may face budget constraints and that impacts their participation [in the WFPC].”
Interview data further describe indirect influence of capacity concerns on representation in the council. Many nonprofits report that they face a dilemma in coordinating with other nonprofits on the council because often these groups are competing for resources with each other (Interviewees 5, 7, and 12). The groups speak of competition less in direct terms but rather are sanguine about the fact that there are simply too few dollars for the number of projects for local nonprofits. Interviewee 12 noted,
And one of the problems, and one of the impediments to it is that we are a bunch of really small organizations involved in this and we are, to some extent, competing for resources. And there just is not a lot of funders for food system work here in the region. And so even if it is widely perceived that it’s in our best interest from a program perspective, it’s not necessarily so from an organizational perspective. So I think that’s the other key issue.
There is also a sense that collaboration is still a valuable commodity even to competing groups:
It’s tricky not to—to get very competitive it feels like. There’s limited resources and limited everything, but “local food”—everybody’s jumping on the local food. So [we are] trying to embrace things with a sense of collaboration rather than competition. (Interviewee 13)
To the extent that the WFPC can enhance nonprofit capacity, or other comparatively low capacity members, it becomes more in the interest of nonprofits to attend meetings. However, like other collaborative governance arrangements addressing boundary spanning problems, the WFPC cannot advocate on behalf of a group nor can the WFPC provide significant funds for group activities. Thus, at this time, the primary benefit the WFPC can offer is a connection to other groups. Given the hyperconnected nature of the region’s food policy system beyond the WFPC, groups lesser known to organizers stand to benefit most from membership. Paradoxically, the WFPC’s original selection process favored actors well known to organizers in the community (Interviewees 6, 7, and 11). Thus, the benefits of increased connectivity were marginal at best for many initial members.
Discussion
Overall, in line with previous work on collaborative governance, our results suggest that group design is important for determining the issues that the governance organization hopes to address (Siddiki et al., 2015). While perhaps an obvious point, it bears repeating that groups can only really be as diverse as they are designed to be. The WFPC exemplifies the approach to collaborative governance which assumes a group’s agenda is specifically intended to be a function of the interests of its members. Thus, representation in the WFPC means ensuring not only that actors are represented but also that issues are represented through membership.
The analysis we report on in this article focuses on the following: (a) verifying the conceptual and practical value of different measures of representation (i.e., descriptive, procedural substantive, and output substantive representation) in a collaborative context, (b) discerning the extent of discrepancy between descriptive and substantive representation, and (c) identifying factors informing observed discrepancies between descriptive and substantive representation. The story that emerges from this analysis follows.
The council is intended to be diverse in its representation by claiming a broad mandate and including a relatively heterogeneous membership structure (OrgType and OrgFoci in the roster column in Figure 3). Data presented in Figure 3 show that, if we think about descriptive representation by presence of group type, we see a small, but important difference between the group as designed and who is literally sitting at the table (OrgType and OrgFoci in the attendance and participation columns in Figure 3). The agenda of meetings is largely driven by an even smaller group of members (contrast OrgType and OrgFoci with Topic in the participate column of Figure 3) who attempt in broad strokes to maintain allegiance to the mission of the organization, but are ultimately narrower in their focus on important issues (trend similarities between OrgType, OrgFoci, and Topic in the significant contributions column of Figure 3).
Our interview data suggest that process norms, lack of structure, unequal resources, and a lack of mission clarity inhibit substantive representation. That is, these factors help explain why representation in substantive ways (i.e., council meeting attendance and agenda issue presence) is generally narrower than representation by design (i.e., membership rosters). Interviewees share that the WFPC is quite open but that local norms perhaps overemphasize process at the expense of material discourse. Interviewees laud the openness of the group but are frustrated by the lack of structure. An open group makes participation easy; however, a structured group makes participation meaningful.
Interviewees also express varying capacity to attend to issues raised by the council, much less attend actual council meetings. Nonprofits have fewer resources and people to contribute to council governance that is peripheral to their overall mission. The effect of capacity heterogeneity in affecting participation in collaboration is documented in previous studies (e.g., Leach, 2006). This factor certainly contributes to the attendance differences we see in Figures 1 and 2, and interviews regarding the relationship between capacity and participation provide a nuanced explanation. Interviewees report that nonprofits compete for the same small pot of funds and collaboration is not everywhere and always viewed more beneficial than going it alone. The capacity arguments contrast with a relatively low cost of commitment to the group (10 meetings a year) and a relatively open agenda space (as evidenced by all interviewees).
Last but not least, interviews reveal a lack of group focus. Mission clarity for collaborative governance arrangements is particularly important (Bryson, Crosby, & Stone, 2015). As the number and diversity of stakeholders increases, so does the number of different goals those stakeholders wish to achieve. Groups with little mission clarity will find either that some representatives will impress their own goals upon the group, that representatives will see their participation as less meaningful, or some combination of both.
The importance of mission is seen further in the evidence that the actors who do understand the WFPCs mission are responsible for driving the agenda of the group. Policy entrepreneurs emerge from two groups: (a) the original founders of the WFPC and (b) representatives from the broader governance structure in which the WFPC is nested. The collection of entrepreneurs forms a core group with a dedicated understanding of the group’s purpose and how the group’s purpose might change. The equitability scores (Figure 4) and interview data suggest that the entrepreneurs are fairly democratic among themselves and cover a range of topics, even though the overall diversity of topics in meetings is less than the group diversity on the council. Policy entrepreneurs are generally respected by interviewees—even complimented—for guiding the group. However, interviewees also expressed anxiety about a future without the entrepreneurs. Thus, the structure-induced diversity in the group design keeps the entrepreneurs honest but cannot, by itself, give the group a consistent policy direction.
Altogether, our research offers critical insights for understanding factors that may contribute to or reduce the gap between descriptive and substantive representation in collaborative governance arrangements. Group norms of openness may contribute to a feeling of an overwhelming agenda, paradoxically leading to less participation by all members. Policy entrepreneurs can dominate meetings but, in this case, can also step in and attempt to represent issues for members who are unable to participate. Notably, group designs must contend with the capacity of members to participate. With this work, we contribute to recent research that seeks to advance an understanding of inclusion, participation, and related issues in collaborative arrangements (Koehler & Koontz, 2008).
Our research findings suggest opportunities and limits for organizational design in fostering substantive representation in collaborative governance. Ultimately, collaboration must be in the interest of nonprofits and other participants to risk expending current and future resources. Thus, collaborative governance arrangements intending to enhance member representation must wrestle with tensions between reducing costs of participation and enhancing payoffs that come from collaboration. Ours is an initial step in thinking about this relationship, and future research ought to investigate this further.
Conclusion: Future Research and Policy Implications
Collaborative governance has emerged as a popular strategy to address issues for which extant policy infrastructure is insufficient (Ansell & Gash, 2008). In many cases, collaborative arrangements are used to address problems for which there are multiple stakeholders, changing conditions, and few governmentally provided resources (Thomson, Perry, & Miller, 2009). As collaborative governance arrangements have grown in number, policy scholars have attempted to identify features of successful organizational structures that contribute to legitimate, effective, and lasting policy (Emerson & Nabatchi, 2015b; Lubell, 2003).
We advance an understanding of the issue of representation in collaborative governance by conceptually and empirically distinguishing representation as descriptive and substantive, and further parsing substantive representation into “procedural substantive representation” and “output or outcome substantive representation.” Much of the literature focused on substantive representation looks at the latter in terms of policy outputs or outcomes. What is missing in current scholarship is attention on the intermediate procedural manifestations of substantive representation that may act as critical precursors to substantive representation modeled as policy outputs or outcomes. Separating the procedural and output or outcome aspects of collaboration to ultimately understand how they relate is typical in collaborative governance frameworks (Emerson et al., 2012; Gerlak & Heikkila, 2011).
Our work suggests that future study of collaborative governance organizations ought to explicitly consider a multidimensional conceptualization of representation as well as a more thoughtful theorizing on the contours of representation. Our study suggests an analytical framework that future researchers might apply to explore variation in group representation, as a function of interaction of design and practice. The conceptualizations of representation and related operationalization we outline here could be applied to any number of collaborative governance arrangements by a wide variety of scholars. We offer a specific strategy of measuring representation in organizational design (descriptive representation) using roster data, organizational representation in practice (procedural substantive representation) using attendance data, and organizational representation in output (output substantive representation) using agenda issue data. The analytical framework that emerges through this research could be used as an explicit tool for practitioners in understanding the functionality of collaborative governance arrangements or by scholars in understanding disconnections between design and outcomes.
Our work contains policy implications for the governance of FPCs. FPCs are spreading across the world as a mechanism by which a variety of stakeholders can address food-related issues. Often, FPCs have little formal authority or capacity to carry out actions; thus, it could be said that the primary purpose of FPCs is to represent previously underrepresented interests. Insofar as the purpose of the council is representation, then, our research speaks to the very core of FPC governance. Councils need to know what they want to do, which issues they wish to cover, and the stakeholders who represent those issues need to feel as though their presence is important in reaching council goals.
Our findings likely confirm what FPCs already know; however, issues of representation are among the thorniest any FPC faces. FPCs are designed such that council representation equals council identity. As such, problems related to issue coverage and other functionality of the council are often framed in terms of council membership. In thinking about ways to create and improve future council actions, the WFPC (like many other FPCs) expanded its membership in 2015. This suggests that the council thinks that membership is a key antecedent to issue representation, a notion that our findings corroborate. Put another way, the expansion of membership suggests that the council is attempting to compensate for the lack of substantive and issues representation but with the wrong strategy. The presence of enough voices on the council is one issue, but getting them heard is the other—and, arguably, the challenge for similar organizations. Councils in a constant search for identity as a means to foster purpose can become stuck looking inward (as in the case of the WFPC at times) or dissolve (as happened to the Portland/Multnomah FPC in 2012). Councils must strike a balance between openness and leadership that ensures that interests are not only represented but that they are integral to collaborative governance (Emerson & Nabatchi, 2015a).
This research contains a number of limitations that influence generalizability. First, the study relies on a single case over a truncated period of time. While we have made every effort to include a broad array of within-case variation, we are ultimately talking about one particular FPC. Insofar as collaborative governance arrangements are necessarily diverse, the extent to which this one case is representative of all cases is limited to the case characteristics we have enumerated. Second, it is likely that a longer term study of the same council would produce more robust findings. Collaborative governance organizations tend to be nimble, particularly those that rely heavily on organizations that must layer their meeting responsibilities on existing tasks. Third, there exists the possibility of selection bias in both the case and the interviewees we have chosen for the case. Because of the small universe of interviewees, obtaining a perfectly representative sample of opinions and interests short of a census is difficult.
Moving forward, we urge future efforts to understand collaborative governance organizations to involve multisite investigations either across governance arrangements that vary within subject (e.g., food), across subject (e.g., food, education, housing), or across governance structure (e.g., nonprofit, government, grassroots). Such studies may yield insights into how collaborative governance arrangements might be designed to address future harms but also to be inclusive of the communities they serve. Furthermore, as collaborative governance arrangements grow as a more popular response to challenging policy problems, the landscape of understanding agenda-setting in institutions may shift from formal governmental sources (e.g., Congress, legislatures, city councils) to governance councils. Studies that examine the agendas of multiple governance structures focusing on similar policy questions might be aggregated to better understand the role of councils in providing an initial locus of problem definition the policy process.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.