Evaluating Social Equity in Protected Areas Co-management: Evidence from the Sundarbans Mangrove Forests, Bangladesh

Abstract Protected areas (PA) co-management can offer equitable benefits from global to local level. Therefore, social equity has come out as a vital goal in PA co-management for ethical and instrumental reasons. In the Bangladesh context, however, social equity issues have been less studied in PA co-management. This study was conducted in the Sundarbans of Bangladesh, and we evaluated social equity as multidimensional concepts of recognitional, procedural and distributional equity. We employed mixed approaches, including 122 structured interviews using a 5-point Likert scale, 15 key informants interview and 4 focus group discussions to collect data on equity perception. Mean and frequency (percent) were calculated and generalized linear models were performed to analyze the predictor’s influence on equity perception. The results showed that recognitional equity was perceived high, while procedural and distributional equity perceived moderate and low. Findings would be useful for policy makers in framing policies towards equitable PA co-management.


Introduction
Across the globe, establishing and managing protected areas (PAs) has become a cornerstone of conserving natural resources as well as securing the livelihoods of local people (DasGupta and Shaw 2017; Kockel et al. 2020;UNEP 2014;Zafra-Calvo and Geldmann 2020). At the 1992 Rio Earth Summit, 150 government leaders agreed to adopt the Convention on Biological Diversity (CBD). The chief instrument of CBD was the Strategic Plan for Biodiversity 2011-2020 which calls for the protection of at least 10% of coastal and marine areas by the year 2020 ensuring effective and equitable management of PAs (Hill, Johnson, and Adamowski 2016;Geldmann et al. 2019). Based on these global efforts, the Government of Bangladesh expanded its protected mangrove forest, particularly three wildlife sanctuaries (the Sundarbans east, west, and south), from 139,699 to 317,950 ha (52% of the entire forest) in the Bangladesh Sundarbans in attention in international conservation arenas (Zafra-Calvo et al. 2017;Franks, Booker, and Roe 2018;Friedman et al. 2018). In the context of PAs, social equity has three interlinked dimensions ( Figure 1): (1) recognitional equity that acknowledges and respects customary rights, traditional cultures and institutions, and traditional knowledge systems of indigenous peoples and local communities in PA governance; (2) procedural equity that implies the inclusive, tenable and fair decision-making in all processes and operations in PA management; and (3) distributional equity that depicts equitable allocation of costs and benefits among stakeholders (Zafra-Calvo et al. 2017;Schreckenberg et al. 2016;Maxwell et al. 2020;Bennett et al. 2020).
The perception of social equity can be influenced not only by social and demographic factors (e.g., age, gender, income, occupation, education, and ethnicity) but also by contextual elements, such as governance, culture, and societal values (Klein et al. 2015;Loft et al. 2017;Wells et al. 2021). Friedman et al. (2020) reported that households with higher income in society were more supportive of forest conservation. Recognitional equity is important, particularly for disadvantaged groups of people . The fundamental rationale of distributional equity relies on the recognition of power dynamics and strong procedures to avoid the elite capture of benefits Halpern et al. 2013). Enhancing social equity in the process and operation requires the use of effective institutional mechanisms (Akamani 2021). Rakotonarivo et al. (2021) reported that the positive perception of social equity could resolve conflicts between biodiversity conservation and other human activities. The presence of social equity in conservation planning and implementation processes can ensure just and effective conservation (Aastrup et al. 2021), while the absence of social equity considerations in the establishment and management of protected areas may also lead to deprivation of local rights (Bennett et al. 2021). Equity in all processes and operations in PA governance may facilitate conservation outcomes (Halpern et al. 2013;Law et al. 2018). Berkes (2004) and Blaustein (2007) have argued that for the success and sustainability of PAs, they must take into consideration social equity issues like recognition of the rights of local people in all spheres. Despite their promise, co-managed PAs have experienced many social equity-related challenges, including unequal distribution of benefits and burdens (Oldekop et al. 2016;Ward, Stringer, and Holmes 2018), less recognition of values, norms and customary rights of local communities (Kabir and Hossain 2008;Dawson et al. 2021), lack of community participation (Damastuti and de Groot 2017), lack of visibility in decision-making processes (Begum et al. 2021), and limited access to PA-related data (Abebe et al. 2020).
Social equity issues in PA co-management have been less studied as compared to other PA governance-related issues like participation, transparency, and accountability (Zafra-Calvo et al. 2017). Dawson, Martin, and Danielsen (2018) assessed social equity in the PA in Laos and concluded that procedural, recognitional, and distributive equity should be considered a PA management goal. Bennett et al. (2020) examined the perceptions of stakeholders regarding social equity in marine PAs in the Mediterranean Sea and reported that small-scale fishermen perceived a positive perception of all dimensions of social equity. Abebe et al. (2020) explored social equity in community-based conservation programs in Ethiopia and reported that economic benefits were not sufficient without recognizing the values of individual identity and promoting fairness in the process of decision-making. Until now, no such initiatives have been undertaken to evaluate social equity as a multidimensional concept in the Bangladesh Sundarban comanaged mangrove PAs. To bridge this research gap, this study was conducted to address the following two objectives: 1. to evaluate the perception of stakeholders about social equity in Sundarbans' PA co-management, and 2. to analyze how predictors (i.e., socio-demographic variables) relate to the social equity perception of stakeholders.
The perception of equity is a powerful factor in human behavior (Loft et al. 2017;Bennett 2016). Understanding the perceptions of stakeholders on equity concerns of establishing and managing PAs is essential to ensuring support for conservation (Aastrup et al. 2021).

Criteria and Indicators
There have been extensive discussions on social equity issues in biodiversity conservation literature (Pascual et al. 2014;McDermott, Mahanty, and Schreckenberg 2013). Perceiving and measuring social equity in PA is critical for achieving effective conservation outcomes (Halpern et al. 2013;Oldekop et al. 2016;Klein et al. 2015;Cetas and Yasu e 2017). Several criteria and indicators have been presented in the literature to evaluate social equity in PA Zafra-Calvo et al. 2017;Franks, Booker, and Roe 2018). Schreckenberg et al. (2016) proposed an equity framework that contains three interlinked dimensions viz., recognition, procedure, and distribution embedded within a set of enabling conditions and sixteen indicators for advancing equity in PA conservation. Zafra-Calvo et al. (2017) proposed several indicators for assessing and monitoring the three dimensions of social equity in PA viz., recognition, procedure, and distribution.
In this study, we used Zafra-Calvo et al. (2017) criteria to evaluate social equity perceptions of stakeholders ( Figure 1) and employed 14 indicators to operationalize the social equity concept in mangrove PAs of the Sundarbans (Table 1). Following Zafra-Calvo et al. (2017) and based on our prior experience, we selected 14 indicators for this study that were location specific and case-dependent as these differ from terrestrial forest ecosystems (Table 1).

Description of the Study Area
The Sundarbans (Figure 2), the world's largest mangrove forests, is shared by both Bangladesh and India (DasGupta and Shaw 2017; Sarker et al. 2016). The area of Bangladesh's Sundarbans is about 6017 km 2 (21 30 0 -22 30 0 N and 89.00 -89 55 0 E), which represents 38% of the country's entire forest area (Iftekhar and Saenger 2008;Hossain et al. 2021). Visibility of decision-making processes and accessibility of relevant information 2.3. Access to justice (AJ) An effective dispute-resolution process 2.4. Accountability Existence of mechanism which makes decisionmakers accountable to all stakeholders and the public 2.5. Legitimacy Acceptance and justification of shared rule by local community 2.6. Free, prior, and informed consent (FPIC) A free, prior, and informed consent (FPIC) obtained from each group of stakeholders affected 3. Distributional equity 3.1. Benefits All the stakeholders receive a fair share of benefits from co-management activities like eco-tourism, non-wood collection, and fishing 3.

Burden
Mitigate burden of local stakeholder groups (e.g., restring access and use of mangrove resources) through co-management actions

Co-management Structure in the Study Area
There are four co-management committees in four forest ranges (Saronkhola, Chandpai, Khulna (Koyra-Dacope) and Sathkhira (Shyamnagar) in the Sundarbans dedicated to governing and managing (Begum et al. 2021). Co-management committees are hierarchically structured (Mollick et al. 2022;Hossain et al. 2021). The Village Conservation Forum (VCF) is formed at the village level and the People's Forum (PF) is formed at the forest range or landscape level. Each PF has an executive committee formed by two members (a man and a woman) from each VCF. The principal committees of co-management include a Co-management General Committee (CMGC) and a Co-management Executive Committee (CMEC), which oversee the overall governance and management of respective PAs (Islam, Ruhanen, and Ritchie 2018). In this study, CMGC and CMEC were together considered Co-

Data Collection
Among the four forest ranges, Khulna (Dacope-Koyra) and Sathkhira (Shyamnagar) were chosen for this study ( Figure 2). We collected a list of 154 VCF from two range offices and chose 122 VCF for this study. The reasons for selecting these VCFs were easy to access and the reliance of these members on mangrove resources for their livelihoods. From each VCF, a representative (usually the chairman of the VCF executive committee) was selected as a respondent for a structured questionnaire survey. The survey questionnaire was designed to generate responses on social equity issues relating to recognitional, procedural, and distributional equity on a 5-point Likert scale (1 ¼ very low, 2 ¼ low, 3 ¼ neutral, 4 ¼ high, and 5 ¼ very high). We conducted an interview with 15 key informants to generate their opinions and experience equity in PA co-management. We also organized four focus group discussions (FGD) with members (3-8) and local resource users from community-based organizations to obtain collective opinions on equity issues in the Sundarbans PA co-management. A checklist was prepared to facilitate KII and FGD questions, such as how the co-management program had recognized the IPLCs customary rights, how the co-management program had shared benefits among the stakeholders, and whether there were restrictions (access to forest resources) imposed by the co-management policy program. Both KII and FGDs were used to triangulate data validity and reliability. The study was carried out between September and December 2020.

Data Analysis
We calculated descriptive statistics including the mean score on the Likert scale, standard deviation, frequencies, and percentages. A correlation matrix among the indices representing each type of equity as well as the socio-demographic variables (age, gender, education, occupation, and income) was developed to observe the strength of the relationships. We performed multiple regressions using generalized linear models (GLM) to analyze the predictor's (age, gender, education, occupation, and income) influence on the perception of the equity of stakeholders. All statistical tests were run using R-programing (R Core Team 2021). We visualized responses to Likert-type data using the 'sjPlot' package (L€ udecke 2021) in R. For qualitative data generated from KII and FGD, a narrative description was followed.

Results
We described the results of three social equity dimensions first followed by overall equity perceptions, as well as the factors that influenced the stakeholders' perception.

Recognitional Equity
For customary rights, 71.3% perceived very high to high (mean score 4.02), while 10% perceived low to very low ( Figure 4 and Table 2). A women key informant commented that "There is no restriction on performing religious rituals before entering into forests for honey and goalpata (Nipa palm) collection. These norms, rules, and practice are not constrained by co-management programs." About 44.2% of the respondents judged traditional knowledge systems as very high to high (mean score 3.33) and 40.2% of respondents were neutral in their responses. For other recognitional equity variables, including traditional culture (mean score 4.19), traditional institutions (mean score 4.20), and women's customary practices (mean score 4.08), the perception of the majority of respondents (70-80%) was very high to high. However, 56.4% of the respondents perceived low to very low livelihood needs (mean score 2.32) (Figure 4 and Table 2). Related to livelihoods, a VCF member commented that "Earlier (before the introduction of co-management in the Sundarbans) we entered into the forests (bada in the local language) for gathering goalpata, and netting fish and fish fry without passes or permits but nowadays we are not allowed to enter into forests without passes/permits. These passes/permits are even given only for a short period of time, but the rest of the time we are workless." Key informants reported that "most of the local people perceived that without adequate alternative income generating activities (AIGA), their lives became miserable." Furthermore, the COVID-19 pandemic aggravated the situation in the study areas by imposing restrictions on entering into the forest. The mean score (i.e., composite index) for recognitional equity was 3.69, which indicates a high level of perception of equity in stakeholders (Table 2).

Procedural Equity
In the case of participation under procedural equity, 41.8% of respondents indicated between very high and high (mean score 3.29) (Figure 4 and Table 2). A member of PF commented that "we have been joining in the general meetings regularly but our voices are hardly recognized." On the other hand, respondents' (68.1%) perception of transparency (mean score 2.0) was low to very low. A similar pattern of perception was observed in the case of accountability (mean score 2.03) where 68% of the respondents perceived it from very low to low. Members of several VCFs stated that "CMEC is not transparent and accountable in the decision-making process, planning, and implementing the annual work plan." Procedural equity in terms of legitimacy (mean score 3.7) was responded from very high to high by 60.6% of the respondents (Figure 4 and Table 2). About 44% of the respondents perceived very high to high for free, prior, and informed consent (FPIC) (mean score 2.67), which indicates that prior consent was obtained from local communities to involve them in PA management activities in the Sundarbans. The mean score (i.e., the composite index) of procedural equity was 2.73, which indicates a moderate level of perception (Table 2).

Distributional Equity
Distributional of benefits to the beneficiaries was a major concern in the Sundarbans PA co-management, which is evident from the respondent's perception. About 80% of the respondents perceived low to very low benefit distribution (mean score 1.77) (Figure 4 and Table 2). The president of CMEC reported that "they have not been receiving the prescribed share percentage (50:50) according to the provision of PA Management Rules, 2017. Members of Community Patrol Groups stated that "we are patrolling jointly with local FD staff to protect the Sundarbans but do not receive any kind of remuneration except a set of uniform, a torch light, and a binocular." PF members stated that "entering into the forests is our customary right to gather nonwood forest products, catch fish and crab, and collect honey from the forests and sell it to the local market, but are now constrained by co-management policy program." The general secretary of CMEC reported that "co-management policy program imposed some restrictions like issuing a limited Boat License Certificate (BLC), a fishing net with a mesh size <15 mm, a fish size limit, poising fishing, a seasonal gear ban (a kind of set bag net), fishing ban during breeding season" which had affected livelihoods of forest-dependent people negatively. The composite index of distributional equity was 2.15, which indicates a lower perception (Table 2).

Association between Socio-Demography and Stakeholders' Perception of Different Social Equity
The results of the linear correlation matrix showed an association between the sociodemographic variables of the respondents and social equity variables (Table 3 and TS1, TS2, TS3). In the case of recognitional equity, traditional institutions showed a significantly negative association with occupation (r ¼ À0.20), while livelihood needs exhibited a significant and positive association with education (r ¼ 0.21) and income (r ¼ 0.37) and significantly negative association with gender (r ¼ À0.32) (Supplementary Material-TS1). Socio-demographic variables had no significant associations with customary rights, traditional knowledge systems, and traditional culture.
Most of the indicators of procedural equity exhibited significant and strong associations. Participation (r ¼ 0.24), transparency (r ¼ 0.27), access to justice (r ¼ 0.30), accountability (r ¼ 0.18), legitimacy (r ¼ 0.28), and FPIC (r ¼ 0.36) showed a significant and positive association with the incomes of respondents (TS2). Education (r ¼ 0.22) was significantly associated with access to justice. Gender showed a significant and negative association with transparency (r ¼ À0.23) and access to justice (r ¼ À0.25) and FPIC (r ¼ À0.18). Legitimacy was negatively associated with occupation (r ¼ À0.07), while FPIC was positively associated with age (r ¼ 0.20). Many of the indicators of procedural equity showed positive and significant associations among themselves (TS2). In the case of distributional equity, there were no significant associations among the indicators of benefits, while burdens exhibited a significantly positive association with income (r ¼ 0.23) and a negative association with gender (r ¼ À0.20) (TS3). Sociodemographic variables, like age, gender, education, and occupation had no significant association with participation and accountability.
While we looked at the association between three criteria (dimensions) of social equity and socio-demographic variables, we found that recognitional equity was significant and positive with income (r ¼ 0.20) and significantly negative with gender (r ¼ À0.18) ( Table 3). Procedural equity exhibited a significant and positive association with age (r ¼ 0.18), education (r ¼ 0.20), and income (r ¼ 0.40), while there was a significant and negative association with gender (r ¼ À0.28). Procedural equity exhibited a significant and positive association with recognitional equity (r ¼ 0.25) and distributional equity (r ¼ 0.55) ( Table 3). Table 3. Linear correlation matrix among socio-demographic variables and the indices of overall equity (recognitional, procedural, and distributional).

Factors Affecting the Social Equity Perception of the Stakeholders
In generalized linear regression, the outcome variables are mean scores of all the indicators under each equity criterion, and the predictors include the socio-demographic factors. The results showed that only income had a significant (p < 0.01) influence on procedural equity and the remaining socio-demographic factors showed no significant influence on the equity criteria (Table 4). Socio-demographic variables had no significant effect on recognitional and distributional equity except income of procedural equity. This indicates that the income of the local people had a significant influence on equity perception.

Discussion
The Sundarbans of Bangladesh have presented a staggering complex due to their diverse socio-ecological systems (Mozumder et al. 2018). The question of equity has garnered attention among scientists, policymakers, and practitioners to secure a sustainable humanosphere in the era of the Anthropocene (Albrecht 1995;Steffen et al. 2015). Even the UN Agenda for 2030 sets equity at the core of sustainable development goals (Leach et al. 2018;Nath, Jashimuddin, and Inoue 2020).

Recognitional Equity
This study demonstrated the positive responses of the respondents toward recognitional equity. Most of the respondents (60%) showed a positive perception toward recognitional equity issues like customary rights, traditional cultures, norms, values, and knowledge of the stakeholders (Figure 4 and Table 2), which corroborate the results of a recent study on social equity in marine protected areas that reported a positive perception of small-scale fishermen toward various indicators of recognitional equity issues (Bennett et al. 2020). Drawing on a global survey of PA management, Zafra-Calvo et al.
(2019) reported an absence of acknowledgment and appreciation for indigenous peoples and local communities (IPLC). It was expected that co-management in the Sundarbans would treat all stakeholders equally and respect their culture, traditional institutions, and the traditional knowledge systems of IPLC. Titumir and Afrin (2018) argued that IPLC contributes significantly and positively to the sustainable management of the Sundarbans in Bangladesh.

Procedural Study
Within procedural equity, the majority of the respondents perceived high in participation of co-management activities (Figure 4 and Table 2). These research findings were supported by the results of Roy (2016) who reported that local community participation in mangrove conservation initiatives was high, but the level of participation was not as encouraging as it was evident from focus group discussion. In the case of participation in general meetings, the physical presence of local stakeholders was high but their participation in decision-making was negligible. Abebe et al. (2020) reported the results of a community-centered conservation initiative in Ethiopia. Martin et al. (2016) argued that the participation of IPLC in making management-related decisions is crucial for success in PA co-management. Respondent's perception toward transparency and accountability was low, which corroborates the findings of Islam, Ruhanen, and Ritchie (2018), who reported that transparency and accountability were low in co-management in the Sundarbans. Begum et al. (2021) reported the lack of transparency in the issuance of forest-entry permits in the Sundarbans of Bangladesh. Results of the correlation matrix show that among all the socio-demographic variables, only the income of the respondents had a significant and positive influence on the perception of procedural equity. Friedman et al. (2020) also reported that households with stable income sources perceived a higher perception of procedural equity associated with community forestry in Indonesia.

Distributional Equity
The majority of the respondents (80%) perceived low distributional equity in terms of benefits ( Figure 4 and Table 2), which is similar to the observation of Abebe et al. (2020). Distribution equity in terms of burdens was also perceived as low, which is similar to the findings of Ward, Stringer, and Holmes (2018) who reported that PA comanagement imposed restrictions, i.e., access to forest resources, which affected the perceived livelihood of the stakeholders. Strict restrictions, such as access to Sundarbans without a permit are burdened on IPLC livelihood (Siddiquee 2020).

Overall Social Equity in the Sundarbans
Recognitional equity was perceived as high among the equity types and distributional equity was perceived as low. Zafra-Calvo et al. (2019) reported significant challenges toward achieving equity in PA management, particularly effective participation in all processes and operations, justice, and the customary rights of IPLC. Procedural equity had a positive correlation with recognitional and distributional equity (Table 3), which reveals that these two dimensions are interlinked. Stakeholders' participation in making management-related decisions is central to the fair sharing of benefits. Without recognition, an individual or group is unable to participate in the community's general meetings (Meerow, Pajouhesh, and Miller 2019). Socio-demographic variables, such as the incomes of the respondents influenced recognitional and procedural equity. Similarly, education has an impact on procedural equity. Gender had a negative effect on procedural equity (Table 3). Hence, the socio-demographic variables of respondents showed differences in perception on different equity dimensions in the studied PA comanagement.

Conclusion
It was expected that the co-management of the Sundarbans PA would treat all stakeholders equally, create a space for all to partake in making management-related decisions, and distribute benefits proportionately. However, the results of this study indicate that social equity in the Sundarbans was impeded by multiple factors including unequal distribution of benefits and restrictions to access to the mangrove resources. Therefore, the PA management plan needs to achieve social equity, ensure the engagement of all stakeholders in designing and executing of the yearly management plan, and distribute benefits according to the mandate of PA Management Rules given in 2017 (Mollick et al. 2022). The study showed that predictors had no significant influence except the income of distributional equity. Social equity needs to be maintained in all processes and operations of co-management policy programs in the Sundarbans. Although the study had some limitations, the empirical study could bring useful information for mangrove policymakers, development agencies, and mangrove managers in maintaining social equity dimensions in co-management policy programs in the Sundarbans.
Considering the findings of this study, we have made the following recommendations to maintain social equity in PA co-management: i. Designing a pragmatic PA co-management program that embraces indigenous people and the local community's customary rights, culture, traditional knowledge systems, and institutions for equitable conservation and human wellbeing. ii. Creating opportunities for the participation of all stakeholders including men, women, youth, traditional mangrove resource users, and the ethnic Munda community in decision-making processes, committee formation, and benefitsharing schemes as it empowers local people. iii. Fostering an enabling environment that helps to engage IPLC in the design, management, and implementation of mangrove PA co-management programs. iv. Maintaining transparency and accountability in all operations and procedures of PA co-management enhances social equity as it entails the openness of the governance system. v. Sharing benefits as per the notion expressed in the existing rules of PA co-management to build mutual trust among the stakeholders. vi. Guarantee local communities' access to justice in conflict situations of mangrove PA co-management. vii. Social equity needs to be considered as an intrinsic management goal to achieve conservation outcomes as local equity consideration can shape the outcome of PA co-management policy programs.

Limitations of the Study
There were 240 VCF in the entire Sundarbans. The study had considered only 122 VCF out of 240 for generating the equity responses of the stakeholders due to the limited access to the mangrove areas as well as financial constraints. We used the indicators in this study that were case-dependent and location-specific. However, the research results had broader implications for protected area managers in the Sundarbans and elsewhere.

Ethical Approval
Informed consent was taken from all participants who participated in interviews. This paper has not been submitted elsewhere for publication.

Funding
This study was funded with the individual pump priming research budget provided by the Faculty of Science and Engineering, University of Nottingham Malaysia to TKN.