Social cohesion and reconciliation between the Rohingya and host communities in Cox's Bazar, Bangladesh

Abstract

Purpose

The Rohingyas are forcefully displaced from Myanmar and sheltered in the Cox's Bazar district of Bangladesh. They have outnumbered the local people indicating a critical condition of their living situation after the year 2017 influx. The aim of this paper is to present how coexistence has impacted social cohesion and reconciliation among different groups of Rohingyas and host communities.

Design/methodology/approach

A cross-sectional survey was conducted with 903 households through a multistage stratified random sampling. Social cohesion and reconciliation (SCORE) index was measured as a multifaceted theoretical construct based on the exploratory and confirmatory factor analyses.

Findings

The findings of this study are inclined toward the miracle of social cohesion and reconciliation between the Rohingya and host communities. Also, it is revealed that youth, males and Rohingyas who came before the year 2017 influx are more cohesive. In the host or local community, people having lower income, less education and involvement with the informal sector are revealed as less cohesive.

Practical implications

This study suggests vocational training as a short-term, cash-for-work as a mid-term and repatriation, proper identity, and protection services as long-term strategic plans to make the two communities more cohesive.

Originality/value

This study focuses on the SCORE indexes with a quantitative format, applying a second-order factor model.

Citation

Miah, A.S.M.J., Islam, T., Fasse, A., Anjum, I., Morshed, K., Akhter, M.E., Sultana, N. and Rayhan, M.I. (2023), "Social cohesion and reconciliation between the Rohingya and host communities in Cox's Bazar, Bangladesh", Southeast Asia: A Multidisciplinary Journal, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/SEAMJ-07-2023-0055

1. Introduction

The Rohingya crisis is one of the most pressing issues in the world. According to UNHCR (2021), in 2017 (25 August onward), the global community witnessed a humanitarian crisis while the Rohingyas of Myanmar were forced to leave their home to take refuge in neighbouring Bangladesh to save their lives. As of 30 September 2021, about 90,2947 Rohingyas (including 190,662 families) live in Ukhiya and Teknaf Upazilas of Cox's Bazar district in Bangladesh. More than 82% of them fled the persecution in Myanmar's Rakhine State and took shelter in Bangladesh, while about 120,000 were born in the camps in the last four years (UNHCR, 2021). Consequently, the population of Rohingyas in Ukhiya and Teknaf has become almost double that of the local Bangladeshis (Bowden, 2018). This demographic alteration indicates the current living conditions and the coping strategies of Rohingyas in Ukhiya and Teknaf. The livelihood of the local host population has been dramatically affected by the incoming Rohingyas in a diverse manner, such as humanitarian concerns and economic and environmental challenges (Grossenbacher, 2020; Ahmed, Simmons, Chowdhury, & Huq, 2021). In particular, the satisfaction of the host community related to social harmony has declined significantly (Biswas et al., 2021). Under these circumstances, one of the main concerns of the protracted crisis is how to improve the existing relationship, particularly social cohesion and reconciliation between the host and Rohingya communities, before their repatriation. There are a few literature on the cohesion in between the refugee and host communities. Maystadt and Verwimp (2014) have studied on the refugee in Kagera, Tanzania (1991-2004) from Burundi, and Rwanda and found that the net economy was emerged in the host area, though a few agriculture-based host suffered for increased competition, surging price of goods. However, non-agriculture and self-employed farmers were benefitted by the refugee. Again Maystadt, Hirvonen, Mabiso, and Vandercasteelen (2019) have concluded that the impact on economy, security and values are unequally distributed among the hosts. Most of the refugees are still located in the world's poorest countries. Access to arms and illicit trade may be of concern. Also disease outbreak, insecurity, environmental degradation, resource competition and price hike were of concern. On the contrary, after the refugee repatriation, the constructed roads and hospitals remain there may lead the hosts to long-term benefits. Tafere (2018) and Maystadt, Mueller, Van Den Hoek, and van Weezel (2020) emphasised on the refugee settlement and environmental degradation in the host area, sometimes forest land would convert to cropland. Taylor et al. (2016) suggested that aid in cash would rather boost up the host economy than the in-kind refugee aid. Their study found that food price was decreased in the local market, for the items distributed in-kind at refugee camp. Also unemployment rate was increased for the low-skilled workforce and decreased their wages. Each additional aid in dollar to the refugees increased income in the local economy by $1.70. While Smith et al. (2021) and Gronau and Ruesink (2021) suggested to enhance refugee and host resource sharing, capacity building programs, community interaction and agriculture ownership to address social cohesion.

Social cohesion gets considerable attention among political decision-makers and international organisations such as the United Nations. It is considered an essential driver in achieving sustainable development goals (UNDP, 2018; IOM, 2020; UNHCR, 2021). Social cohesion is defined as the civic engagement and institutional association of the Rohingya and host communities, and reconciliation is defined as the interaction between the Rohingya and host communities (UNDP, 2015). The indicators of a higher social cohesion and reconciliation (SCORE) index are healthy, positive and integrated relationships. Thus, the SCORE index is a multifaceted issue. The SCORE index is used in Cyprus, Bosnia and Nepal to understand the dynamics of affiliation among the different communities (UNDP, 2015). The SCORE index also requires primary data with in-depth multidimensional indicators. For facilitating social cohesion among the Rohingya and host communities, it is crucial to focus on conflict and the pattern of coexistence. Towards that end, the SCORE index and its determining factors, such as trust, identity, equality, access to resources and livelihoods, are yet to be analysed relatively.

Given the complexity and challenges of hosting many Rohingyas in Cox's Bazar district in Bangladesh, it is crucial to explore better ways to support the needs of Rohingya and host communities in different terms. That is why it is necessary to look into possible windows of opportunity. This study is thus striving to unveil the current relationship, conflicting issues and new avenues of initiative; a conceptual framework is mentioned in Figure S1. The primary research question is whether the Rohingya and host communities facilitated cohesion and reconciliation as a coping strategy and, if so, to what extent?

The sections are arranged in this study as follows: a brief introduction in the first section, and the subsequent section is organised with the context and abridge history behind the community cohesion. The context follows the methodology with a detailed survey design, data collection tools and protocols. A few referred methods are discussed in the methodology to explore the analytical framework. The following section illustrates the analysis and results. The last two sections include the discussion and conclusive remarks, which ended up with the references.

2. Methodology

2.1 Study design and respondents

From 10 to 25 December 2020, a cross-sectional survey was conducted in the Ukhiya and Teknaf Upazilas of Cox's Bazar district among the adult Rohingyas (before and after the 2017 influx) and host communities in Bangladesh. The Rohingyas live in two registered camps and 32 unregistered makeshift camps, where Kutupalong (in Ukhiya) has the largest concentration of displaced households (UNHCR, 2021).

The sample size of Rohingyas was more in numbers of newly arrived Rohingyas than the Rohingyas who lived before 2017. In total, a survey with 483 Rohingya households (367 from ‘after the 2017 influx' and 116 from ‘before the 2017 influx' groups) was done. Similarly, a survey with 420 households from the host community of two Upazilas (198 from Ukhiya and 222 from Teknaf) was conducted. A two-stage stratified random sampling (Upazilas - Camps - Households) technique was followed to select 903 households. The required sample size was calculated at 80% power, at 95% confidence interval, considering the proportion of cohesion to be 50%, design effect was considered as 2 and with 10% non-response rate. The survey was conducted with the voluntary participation of respondents. A face-to-face interview took place, maintaining proper social distancing. The questions for the social cohesion and reconciliation score were arranged in 5-point Likert scale: strongly agree, agree, neutral, disagree and strongly disagree.

2.2 Methods

SCORE index is measured as a multifaceted theoretical construct, applying a second-order factor model (DeYoung, Peterson, & Higgins, 2002; Hills and Argyle, 2002). From the observable variables, the model estimates a first-order latent variable called indicator and their respective factor loadings through exploratory factor analysis (EFA) (Johnson and Wichern, 2014). The internal consistency of each loading is tested using Cronbach's alpha coefficient with a threshold over 0.8 (Field, 2005). Once the factors are identified through the EFA, second-order latent components are measured through confirmatory factor analysis (CFA), considering the significance, strength and direction of factor loadings (Johnson and Wichern, 2014). Another sensitivity analysis, first-order stochastic dominance property, is performed for precision tests (Haughton and Khandker, 2009). The next step is to examine how items are loaded onto indicators and indicators loaded onto their components. According to the UNDP (2015), the SCORE index ranging from 0 to 10 is being estimated through the factor loadings as follows:

• Step 1:

The weighted score (factor loading) is measured for each indicator (component) from the observed variables (items) using equation (1):

(1)Wα=w1x1+w2x2+w3x3+.........+wkxk

where Wα is the weighted score (factor loading) of the first-order component using CFA, wi is the loading of the corresponding i th variable, xi is the observed i th variable or response from the questionnaire, k is the number of variables used in the model.

• Step 2:

The next step is to rescale the Wα to a range of 0 to 10, to do so, the following two equations are measured:

(2)Wmα=w1xm1+w2xm2+w3xm3+.........+wkxmk

where Wmα is the theoretical maximum weighted score of the component, wi is the loading of the corresponding i th variable onto the component, xmi is the maximum value of i th variable, k is the number of variables used in the model.

(3)Wrα=Wα *10Wmα

where Wrα is the rescaled weighted score of a component between 0 to 10 and the values of Wα and Wmα are found in equations (1 and 2), respectively.

• Step 3:

Steps 1–2 are followed for each of the components. The following equation thus computes the weighted score of each component:

(4)Wβ=v1wrα1+v2wrα2+v3wrα3+.........+vpwrαp

where Wβ is the weighted score (factor loading) of the second-order global dimension (cohesion and reconciliation) using the CFA, vi is the loading of the corresponding i th component, wrαi is the rescaled weighted score of i th component, p is the number of components (factors found from the EFA) used in the model.

• Step 4:

The same steps are followed here from step 2 to rescale the global dimension:

(5)Wmβ=v1wmα1+v2wmα2+v3wmα3+.........+vpwmαp

where Wmβ is the theoretical maximum weighted score of the second-order dimension, vi is the loading of the corresponding i th component, wmαi is the maximum value of i th component, which would be 10, p is the number of components used in the model.

(6)Wrβ=Wβ *10Wmβ

where Wrβ is the rescaled weighted score of dimension between 0 to 10, and the values of Wβ and Wmβ are found in equations (4 and 5), respectively.

Here, Wrβ value after the four steps provided the scores of social cohesion and reconciliation. Meanwhile, CFA includes a maximum likelihood (ML) algorithm to estimate the loadings, and 10000 bootstrap iterations have been taken for the convergence (Johnson and Wichern, 2014).

3. Analysis and results

The overall SCORE index (as indicated in Figure S2) for the Rohingya community is higher than that of the host community as their cohesion with institutions is much higher than that of the host. Rohingya people are mostly aid-dependent for their survival, living with an uncertain future in the camps vulnerable to monsoon menaces. They are stripped of any national identity, and most are not even recognised as “refugees” (Ahmed et al., 2021).

Figure S2 depicts that host community has 1.45 less SCORE than that of Rohingya community, as forcibly displaced nationals and part of their coping strategies Rohingya people are more cohesive towards the government and non-governmental institutes.

Figure S3 delineates the social cohesion of the Rohingya community from a CFA. Social cohesion is indexed by the factors: human security, satisfaction with humanitarian aid, trust in government officials and trust in institutions (UNDP, 2015). Variables confounded in each factor are classified by the loadings of EFA with a high Cronbach's alpha (0.82). The correlation values between the factor and variables are shown in the arrow marks. The strength of trust in institution possesses the highest correlation (0.76) to address institutional cohesion. Among the variables, the national NGOs and international NGOs depict the highest correlation. The relationship between the Rohingyas and the host community has been greatly impacted by the humanitarian aid provided by both national and international NGOs. The second best factor for social cohesion is found to be the trust in government officials (0.57), among the variables trust on union chairman (local community leader) shows the highest correlation. Human security factor is correlated with the safety net programs in major, whereas satisfaction with humanitarian aid is mostly associated with the aid workers amiable attitude (0.86).

Figure S4 shows the intercommunity reconciliation with factors: positive feelings, cultural similarities, intergroup contact and propensity for retribution. CFA analyses the relationships between the factors and variables or indicators. Positive feelings (0.69) show the highest correlation factor for community-based reconciliation. Among the variables of positive feelings, sympathy of Rohingya community towards the host community reveals the highest correlation (0.82); also respect and affection are very close and high association (0.81 and 0.79 respectively) with the positive feelings in between Rohingya and host communities. Cultural similarities are revealed as the second highest factor (0.63) associated with the intercommunity reconciliation. Similarities of language and behavioural norms in between Rohingya and local communities are figured out to be highly associated. Intergroup contact is defined by the two variables in major, such as host attends Rohingya's social events and Rohingya frequently invites the host in their events. Overall goodness-of-fit test: root mean squared error of approximation (RMSEA) p-value< 0.05, which means CFA is statistically significant.

Table S1 depicts the stories behind the Rohingya community's cohesion and reconciliation in a segregated way. For the overall SCORE index, old aged people dominate over the youth. Teknaf shows a better scenario than Ukhiya. Educated one (class 5+) shows upper value. Male respondents possess a higher index than females. The cohesion index with institutions is higher among the youth Rohingya, but older people show better community-based reconciliation. People living in Ukhiya show higher cohesive relationships with the civic society and institutions but a lower reconciliation with the host community. Most Rohingyas were deprived of education in Myanmar (Shohel, 2022). However, those with primary education or more strongly associated with the local community. Female Rohingyas are more cohesive towards institutions, while males have higher reconciliation with the host community. With the rise of income, cohesion or association with the institutes decreases and intercommunity reconciliation increases. It can be concluded that economically and socially vulnerable people are more inclined to have institutional cohesion rather than community-based reconciliation.

For the Rohingya community, a sensitivity post-estimate analysis (first-order stochastic dominance) also supports the findings of Table S1. Figure S5 delineates that the SCORE index is dominated by males against females and Teknaf over the Ukhiya. First-order dominance refers to a concept that compares two cumulative distribution functions. First-order dominance is important because it implies that any social welfare function that is increasing in nature will record higher levels of welfare in the dominant factor. Female Rohingya community shows less than a 0.5 probability, means higher welfare or SCORE index remains for the male counterpart. Again for Rohingya community living in Ukhiya has a 0.396 probability of dominance, means Rohingya people living in Teknaf have higher dominance in SCORE index.

Figure S6 shows the distribution of per capita income and expenditure over the indicator variables. In Teknaf, the income and per capita expenditure are higher for the Rohingya community. The larger mean and variation depict a better economic environment. Again those who came before the 2017 influx have superior financial status and job opportunities. Many of them also received institutional identity. Male-headed households from the Rohingya community expressed a stronger earning ability than the female-headed counterpart, but the variation of expenditure is higher for female respondents. Youth have substantially superior earning opportunities compared to older Rohingyas.

Table S2 shows the policy-oriented inequality measure (Gini index). A general hypothesis is that a higher income generates higher inequality. The Rohingyas living in Ukhiya and arriving after 2017 depict a lower income with marginally higher inequality. Those who came before the 2017 influx mostly live in Teknaf and have greater engagement with formal livelihood activities, business and better community relations with the hosts (Uddin, 2021).

Figure S7 displays the social cohesion of the host community with the institutions. EFA confirms the clusters of variables with the factors, and a high Cronbach's alpha was found (0.77). The highest dominating factor is the trust in institutions (0.72) to address the cohesion. Both the local NGOs and international NGOs show high correlation (0.97 and 0.94 respectively) with the institutional trust factor. The next associated factor for social cohesion of the host community is the trust in government officials (0.63), and the most influential and correlated variable is found to be union chairman (0.92). Human security becomes the third most highly associated factor (0.49) for social cohesion in the host community, and less crime incidents is extracted as the most related variable with this factor. Figure S8 reveals that intergroup contact is the highest correlated (0.65) factor to explain the host-Rohingya intercommunity reconciliation. Among the variables, host attends the Rohingya's social event (0.91) and host invites the Rohingya's (0.85) disclose as the two most correlated variables with intergroup contact factor. The next most influential factor is found to be the positive feelings (0.61) and affection of host community towards the Rohingya community is revealed as the most correlated variable (0.93). Cultural similarities evolve as the third influential factor, and behavioural norms (0.78) is the most associated variable to explain this factor. A CFA is performed to get the correlation matrix and the factor loadings after a detailed EFA with Cronbach's alpha of 0.81. Also, the easing of intergroup anxiety and negative stereotypes are the significant factors to address reconciliation. Overall goodness-of-fit test: root mean squared error of approximation (RMSEA) shows the p-value< 0.05 which means CFA is statistically significant.

Table S3 illustrates the social cohesion and reconciliation of the host community respectively with the institution and the Rohingyas. Regarding the overall SCORE index, male, youth and less educated hosts possess higher values than their counterparts. Although the youth hosts have a higher reconciliation index with the Rohingyas, they have a lower cohesion index with the institutions. Youth and male hosts have more reconciliation with the Rohingya community as they engage with Rohingyas through work, business, marriage, sports and other online–offline social communications (Uddin, 2021).

The Ukhiya host community exhibits a better reconciliation with the Rohingya community and cohesion with the pertinent institutions than Teknaf. The majority of the Ukhiya Rohingyas live in camps. Hence, job and business opportunities are higher in Ukhiya. However, the host community in Ukhiya has become vulnerable as they are a minority demographically against the Rohingyas. They are the most sufferers of the recent Rohingya influx (Mora and Yousuf, 2021; Olney, Badiuzzaman, & Hoque, 2019). Hence, their vulnerabilities and sufferings have led them to cooperate more with the institutions and authorities concerned with the Rohingya rehabilitation program.

For the host community, a sensitivity post-estimate analysis (first-order stochastic dominance) is performed in Figure S9. Males dominate over females in the SCORE index. And the host community in Ukhiya delineates a higher SCORE index than the people from Teknaf. Probability score for female over male is below 0.5, and hence, a lower dominance is observed, whereas the probability score for the host community in Ukhiya shows a much higher propensity over the host community lives in Teknaf.

Figure S10 shows the per capita income and expenditure distribution over the age, area, gender, education and profession of the host community. Youth in the host community has a higher per capita income and expenditure than the older-aged respondents. Local people living in Ukhiya possess a better economy than the people of Teknaf. Hosts with above primary education have higher income, and formal job holders show better earnings and expenditure.

In Table S4, the Gini index values of different categories of the host community reveal greater inequality among females than their male counterparts, likely caused by the persistent gender wage gap in the country and locality. The youth and more educated host populations also have greater income inequality than their counterparts, as unemployment is more persistent among these socioeconomic groups. Ukhiya has higher income inequality compared to Teknaf. Also, greater inequality prevails among the hosts engaged in business and formal jobs.

4. Discussion

Social cohesion and reconciliation among the Rohingyas and host communities depend on the condition of the Rohingyas in host societies and the extent and degree of trust, acceptance, interactions, reciprocity and the relationship between communities and the institutions. Moreover, the degree of social cohesion and reconciliation varies with location, power structure and gender.

The overall SCORE index for the Rohingya community is higher than that of the host community as their cohesion with institutions is much higher than that of the host. The host community feels aggrieved sheltering the Rohingyas since the latter's arrival, and protracted staying in Cox's Bazar caused economic difficulties, sociocultural deterioration and environmental degradation and posed an “existential threat” to the former's safety and security (UNDP, 2018; IOM, 2020). The daily life of the host people has been severely disrupted as the price of everyday commodities has sky-high by the crowd pressure in the markets and roads and by mounting traffic congestion (Olney et al., 2019). Children's education has been hampered severely as teachers increasingly opted out for better-paid jobs. Students are also not attending classes to avail themselves of small-job opportunities or for even fear of safety and security (Olney et al., 2019; Grossenbacher, 2020). In addition, there has been severe degradation of the environment and forest land (Mora and Yousuf, 2021) and rapid growth of extremisms and other crimes, including stealing, robbery, human and drug trafficking, harassment, violence and conflict in recent times (Ahmed et al., 2009). These negative impacts of the influx made the locals feel that they have been punished for their generosity and hospitality. Their frustrations are aggravated with their feelings of being excluded from the humanitarian response and apprehending that the NGOs are providing aid disproportionately, even to the “the rich” Rohingyas and are providing lucrative jobs to the “outsiders” depriving the local host people (Ahmed et al., 2021; Grossenbacher, 2020). All the factors mentioned above increased the frustration and anger of the host community in place of empathy towards the Rohingya community. Therefore, the growing hostility of the host community impacted their SCORE index as they have a lower index value (overall 5.30) in both social cohesion index (5.36), which is related to their relations with community institutions, and in reconciliation index (5.41), which is related to their relations with the community people including Rohingya community.

5. Conclusion

The study shed light on the current relationship, conflicting issues and mitigation strategies by analysing the SCORE index between the Rohingya and the host communities. Rohingyas, those who came before the 2017 influx, are in better condition. So are the youths and the males, particularly of both influxes. Therefore, a scrutinised policy is needed to uphold the economic and social standards for the aged, female, low-income groups and Rohingyas who arrived after the 2017 influx. In the host community, those involved in agriculture and informal sectors, i.e. people from low-income groups, aged, less educated and living in Teknaf are getting lesser economic opportunities. Focused vocational training and skill development for the young people in both Rohingya and host communities might be an effective remedial policy. Cash-for-work for the low-income groups in host communities may be a good option to minimise their grievance over the impact on their lives and livelihoods caused by the Rohingya influx.

Inclusive programs that pay equal attention to both communities can help reduce tensions and build a tolerant and stable society. Existing social networks and structures of the host community should be positively utilised to encourage active participation and involvement. Community groups and institutions can play a crucial role in minimising the tensions and increasing the cohesion between Rohingyas and host communities. They can be used as agent of change to spread peace among the people in both Rohingya and host communities. They can also be the gatekeepers to protect their communities from illegal activities or criminal activities that monger social tensions and create hatred towards the Rohingyas. Capacity-building activities should be included in the initiatives to enable the social and institutional leaders and volunteers to work on social cohesion and peace among the Rohingya and host communities. Moreover, the Rohingya crisis response projects should be community-based and coordinated through a flexible approach that considers humanitarian, developmental and peace-building aspects till their repatriation.

Supplementary material

The supplementary material for this article can be found online.