1 Introduction

Bangladesh is highly vulnerable to natural hazards due to its flat topography, number of large rivers, location in the cyclone zone, and climate change (Dasgupta et al. 2014; Mahmud and Prowse 2012). Coastal areas of Bangladesh experience frequent, severe cyclones, bringing wind damage, storm surges, flooding, and saline intrusion (Paul et al. 2012; Rotberg 2010). These hazards, as a whole, create socio-economic and ecological vulnerabilities, and the relative poverty of coastal inhabitants worsens the vulnerability of these communities to such disasters (Alam and Rahman 2014).

To address these vulnerabilities, local people use various networking relationships. Local social capital resources and networks are a vital contributor to community resilience and recovery from disasters (Hawkins and Maurer 2010; LaLone 2012; Minamoto 2010). Isolated households with small social networks are especially vulnerable to disaster, and they receive less organisational support and experience a slower pace of recovery than households with strong social networks (Tse et al. 2013). Aldrich (2010) shows social capital plays a pivotal role in recovery, and specifically that local initiatives and interventions play a key role. Islam and Walkerden (2014) found that immediately after a disaster, affected households depend on bonding networks (relationships with immediate family members and relatives) and bridging networks (relationships with neighbours and friends) to cope with crises, because organisations that provide support take at least a few days to reach the affected area. The capacity of these bonding and bridging relationships is limited by the households’ physical and financial capital, the strengths and weaknesses of these horizontal relationships, and the magnitude of the disaster. For longer-term recovery, disaster victims usually need support through linking social networks, for example, from local government, NGOs, and other community-based organisations (CBOs).

In Bangladesh, there are three layers of local government: Union Parishad, and two layers above it, Upazila (sub-district) Parishad and District Parishad. Union Parishad (UP), the most local form of local government in Bangladesh, is an important medium through which households are linked to recovery resources from outside the affected area after a disaster. UP maintains law and order, implements development schemes, and provides public welfare services at the grassroots level. This paper discusses the supporting role of Union Parishad and documents the process by which UP contributed during the Cyclone Sidr recovery process.

The elected body of UP comprises a Chairman and 12 Members, one for each of nine Wards (a Ward is an electoral unit of a Union Parishad), and three women members each representing three Wards, who are also elected through direct vote. The administrative staff comprises a UP Secretary and dafadar (village police), and nine chowkidars (village watchmen) who are recruited by the UP Chairman (Asian Development Bank (ADB) 2004). The funds UP has at its disposal are modest. They include taxes, tolls, fees, and other charges imposed by the UP under the Local Government (Union Parishads) Act 2009, and rent from UP owned property, grants received, and profits from investments. The UP representatives (Chairman and Members) work full time and receive honorarium from the government. Upazila level political party leaders (e.g. Member of Parliament and Upazila Chairman) dominate selection of candidates for UP elections. Prior to March 2016, candidates were not officially identified with political parties, although they were in practice their representatives. From March 2016, the Election Commission of Bangladesh has formally introduced direct participation of political parties in the Union Parishad election.

The literature on the role of local government in disaster recovery is relatively limited, and there is, to our knowledge, none specifically on the role of local government in Bangladesh. This study contributes to the literature on local government’s roles in disaster recovery and makes an initial contribution to the discussion of its role in Bangladesh specifically. It examines when UP performs well and poorly in disaster resilience and recovery; what factors are responsible for weak performance; and the opportunities for increasing the UP’s capacity to actively contribute to disaster resilience and recovery. It addresses these research questions by a micro-level investigation of how the relationships between households and UP promote disaster recovery, using as case studies two villages affected by the devastating Cyclone Sidr.

2 Methods

2.1 Study area

This study was conducted in South Charduani and Tafalbaria villages of Charduani Union (the smallest local Government unit), Pathorhgata Upazila of Barguna district between February and July 2013. The villages are situated on the Southern coast of Bangladesh, on the banks of the Baleshwar River—a channel of the Bay of Bengal (Fig. 1)—where a significant number of households exist close to and/or outside the protection of the embankments. Study villages were selected purposively based on earlier studies which indicated that the coastal villages are susceptible to multiple hazards (Paul and Routray 2013). The study villages were selected for several reasons:

Fig. 1
figure 1

Source: Banglapedia 2006

Study villages.

Full size image

  • They are located within a high cyclone risk zone (Fig. 2) and are highly prone to cyclones and storm surges. Six major cyclones—Sidr 2007; Reshmi 2008; Nisha 2008; Bijli 2009; Aila 2009; and Mahasen 2013—hit this area (and our study villages in particular) in the seven years from 2007 to 2013 (Bangladesh Meteorological Department 2013).

    Fig. 2
    figure 2

    Source: CEGIS, 2008

    Cyclone tracks.

    Full size image

  • Cyclone Sidr severely affected these villages, causing 344 deaths of inhabitants, which was 4 % of the total population of the villages; about 90 % of houses in these villages were destroyed (Multi Task 2009).

  • Consequently, the villagers’ responses to Cyclone Sidr reveal aspects of resilience and recovery in extreme circumstances, which we expected to be useful for developing cyclone disaster management policies for the Bangladesh coast.

2.2 Study approach

A mixed method approach, using both quantitative and qualitative data from various data sources, was employed to investigate how Bangladeshi coastal households used their linking networks in their recovery process. Using mixed methods facilitates data triangulation (Creswell and Clark 2007). Numerical data were used to identify broad patterns in social life—specifically in matters relevant to recovery processes—in the study villages. Qualitative data were analysed to develop richer descriptions of local perspectives, for example, regarding trust, peoples’ experiences with cyclones, their needs, and other resources that households used for post-disaster recovery and resilience (Chamlee-Wright and Storr 2011).

2.3 Sampling

The study used a combination of random and purposive sampling (Chamlee-Wright and Storr 2011). The study villages were selected purposively: we identified a disaster-affected coastal district, then an Upazila, a Union, and finally, two villages, ensuring that the study villages were typical in the region, and had been substantially impacted by Cyclone Sidr. We interviewed 159 randomly selected household-heads and obtained valid responses from 156 (74 from South Charduani, and 82 from Tafalbaria), sufficient for 93 % confidence that the percentages observed in our sample are within 7 % of those typical of households in the villages (Yamane 1973). (A 95 % confidence interval would have doubled the level of fieldwork required, and 93 % confidence was considered to be sufficient to inform policy development, given that our approach hinges on triangulating multiple lines of evidence).

2.4 Data sources and data collection tools

The first author and two research assistants were directly involved in conducting face-to-face interviews of household-heads using a structured questionnaire with 43 closed- and open-ended questions. Four in-depth case studies were conducted to understand individuals’ experiences in recovering from Cyclone Sidr. Semi-structured checklists were used to guide conversations in a number of settings. Eight focus group discussions (FGDs) with the villagers were conducted to explore how affected households perceived the threats of cyclones, faced the recovery challenges, and benefited from the involvement of UP. Thirty-seven key informant interviews (KIIs) took place with local leaders, local government officials, and village headmen, and workshops were held to discuss issues with journalists, local government staff (chowkidars), and NGO officials. Interviews were also conducted with five policymakers (executives of ministries and departments—policy maker interviews (PMIs)), and 14 disaster practitioners [national and international NGO officials and freelance researchers—disaster practitioner interviews (DPIs)] to explore the outlooks of national-level disaster experts. The participants in case studies, FGDs, local KIIs, and PMIs, were selected purposively, and a snowball sampling technique was followed to identify the participants for DPIs. Notes were taken during interviews, workshops and other group processes. In addition, field observations were used to assist with understanding the physical settings and everyday practices of village life (Pouliotte et al. 2009). During the interview sessions, questions were asked about the effectiveness of UP initiatives in rebuilding the community and returning it to “normal” after Cyclone Sidr, and how UPs’ capacities could be strengthened to effectively contribute to disaster resilience and recovery.

2.5 Data analysis

The quantitative data were analysed with Statistical Package for Social Sciences (SPSS) which provided descriptive statistics, for example, frequency distributions, percentages, and tables (George 2008; Paul and Routray 2013). Qualitative data were analysed by coding and grouping to identify key themes (Islam and Walkerden 2014; Schutt 2011). Qualitative and quantitative data were considered together, to increase confidence in interpretations, and to provide a combination of breadth and depth.

2.6 Ethical considerations

All study protocols were approved by Macquarie University’s Human Research Ethics Committee (reference number 5201200877) and were designed to minimise risks from disclosure of views about corruption. The researchers only raised questions about possible corruption in private/individual discussions (e.g. household survey and key informant interviews). When corruption was raised in focus groups and people started to point out individuals, the first author and research assistants intervened and said the focus groups were not an appropriate medium for that conversation. Sources of views on corruption are not identified personally, to protect their anonymity.

3 Performance of UP in disaster recovery

3.1 Post-Sidr support provided by UP

Ninety-four percent of household-heads reported that UP provided them with emergency relief, in the form of rice, pulses, cooking oil, puffed rice, water, medicine, etc. (Table 1). UP also provided cash and building materials for use repairing partially damaged houses (Table 1, FGDs, and local KIIs). FGDs and DPIs mentioned that UP also provided some new houses, mostly using donations from foreign governments. Most of the relief goods that UP distributes are received from the national government after a disaster. Sometimes, national and international NGOs also work in collaboration with UP to distribute relief amongst the disaster-affected households.

Table 1 Types of UP support after Sidr.
Full size table

Restoring livelihoods is an important form of support after disasters. UP played a much smaller role in this area. Only 12 % of household-heads were aware that UP made contributions to livelihood support, by, for example, distributing crop seeds and fishing nets, and funding repairs for fishing boats (Table 1). FGDs reported that NGOs provided much higher support for re-establishing livelihoods than UP (Islam and Walkerden 2015).

Despite reconstruction after disasters being a responsibility of local government under the central Annual Development Programmes, limited contributions to reconstruction from UP, were evident after Sidr. FGDs and local KIIs advised that the reconstruction works were mostly done by the NGOs, and that they had sought help from international organisations. During field observations, billboards of national and international NGOs (e.g. BRAC, USAID, Save the Children, and South Asia Partnership-Bangladesh) were seen at many road and culvert reconstruction sites. Only 26 % of household-heads were aware that UP was involved with reconstruction of roads and culverts (Table 1).

Post-disaster reconstruction is one of the sectors experiencing major corruption. For UP representatives, this sector is an important opportunity to misappropriate funds, as cash flows are relatively large here after disasters (Mahmud and Prowse 2012, Table 2, FGDs, DPIs, the NGO workshops, and meeting with journalists). Focus groups claimed, for example, that UP did so-called reconstruction works by placing mud on damaged roads under the national government’s “Cash for Works” programme, with most of the money being pocketed. The UP representatives said that the Government allocation of funds to UP was small in the light of the damage that Sidr caused. Consequently, it was not possible for UP to reconstruct many damaged roads and culverts. Reconstruction of embankments was not included in UP’s activities, as these are maintained by the Water Development Board (WDB).

Table 2 Post-Sidr corruption identified by household-heads.
Full size table

In addition, UP made a limited contribution to search and rescue immediately after the disaster (only 15 % of households identified contributions in this area). This is not surprising, given that local UP employees are, in most cases, victims in disasters as large as Cyclone Sidr, and connections with local offices of UP are disrupted initially (with damage to roads, telecommunications infrastructure, etc.). They are also partly constrained by logistical issues, for example, a lack, at the village level, of axes, saws for use in clearing fallen trees off roads to enable search and rescue operations to proceed.

Overall, this study found that UP post-disaster recovery works were relief-centred. Long-term recovery activities—for example, reinstating livelihoods, disaster risk reduction (DRR) education, prevention, preparedness, and employment creation—were given less priority by UP. Both are essential for community resilience, of course.

3.2 Favouritism and uneven distribution of relief and other recovery support

After Sidr, UP representatives favoured their political partymen (political party members and other supporters), and their family, relatives and friends, during distribution of relief goods and other recovery support, for example, distribution of building materials, cash, cropping seeds, Jatka Footnote 1 rice, and VGFFootnote 2 (Vulnerable Group Feeding) cards (household survey, FGDs, KIIs). Focus groups reported that due to favouritism, many non-farming households received cropping seeds as a post-disaster livelihood support, depriving some farming households of the opportunity to rebuild their farms. One household-head claimed that “some people got such quantity of relief goods, and they didn’t need to buy anything for a long time after Sidr, while some people received very little”.

Political favouritism was high during relief distribution. The personal political interests of middlemen, village political leaders, and village headmen, and party politics, led UP representatives to favour their political partymen. The UP representatives are highly involved with party politics; they cannot ignore the demands of partymen, because the partymen work hard to win them votes in the election. One household-head claimed that “I didn’t get all of the relief items that UP was distributing, as my political ideology was different from the current Chairman”. Another household-head said that “the existing Chairman doesn’t like me, as I am a supporter of the ex-Chairman—his opponent. Therefore, I received very little relief goods”. A post-Sidr report from a daily newspaper also found evidence of UP representatives’ favouritism and political bias in relief operations. It reported that UP representatives gave priority to their own supporters in preparing lists of affected people for relief distribution (New Age 2007). Mallick et al. (2009) found misappropriation of UP’s relief works after Sidr and argued that this misappropriation weakens community resilience, increases conflicts, and deprives the poor. Favouritism weakens bonds between UP representatives and the households not aligned with them, while, in certain respects, strengthening bonds with those who are politically aligned. Even for the latter, however, favouritism has a problematic aspect, as political power can shift. Reducing corruption was desired by most (household survey). Accusing the village headmen, FGDs reported that when relief supplies came to the UP office, the UP representatives usually informed the village headmen and village political leaders. The village political leaders rarely informed the common people about available relief; rather they distributed relief among their partymen, friends, and relatives. Eighty-four percent of household-heads said that because of favouring the partymen, UP representatives failed to perform well after Sidr (household survey).

UP Chairmen also often favour their own village when providing recovery support after disasters. A case of favouring one’s own villagers was found regarding infrastructure rebuilding after Sidr. For example, a UP Chairman built substantially more abashon (government-funded community housing for the destitute) in his own village than the needs of nearby villages warranted (FGDs).

Jatka rice is supposed to be distributed only to fishers, and they are to be identified by checking whether a household has nets and boats. But non-fishing people received some of this support, not fishermen. One local key informant said that “real fishers didn’t receive jatka rice—those who received it, their primary occupation was not fishing”. One household-head commented that “the UP Member didn’t give us a VGF card and even pocketed the jatka rice”. FGDs reported that if they visited the UP representatives’ homes, they could get full cartons of relief goods.

As is commonly the case in Bangladesh (Khan and Islam 1991), the UP Chairman often favoured his own villagers to provide recovery support. The villagers of South Charduani claimed that the Chairman distributed more support in his own village, Tafalbaria, than to them. They saw that most of the affected households of Tafalbaria received new houses, whereas many of those affected in South Charduani did not (FGDs). Ninety-three percent of household-heads said that due to this favouritism, UP representatives do not play a neutral role during recovery (household survey).

3.3 Corruption during relief distribution

UP representatives often received bribes during distribution of recovery support. Most of the national government’s recovery support is distributed through UP. Ninety-two percent of household-heads reported corruption in relief goods distribution; 72 % reported corruption in cash distribution (Table 2). Eighty-five percent of household-heads identified that some UP representatives were corrupt (Table 2). One disaster practitioner claimed that after Cyclone Aila, a UP Member deposited Tk. 1,500,000 (US$ 19,480, US$1 = Tk. 77 as of April 2013) in the bank. This is a large amount in the context of rural Bangladesh. FGDs reported that due to involvement with post-Sidr corruption an ex-UP Member of South Charduani could not contest the last election. The UP also distributed cash as a Government contribution to repairing damaged houses. The households with fully and partially damaged houses received Tk. 5000 (US$ 65) and 2500 (US$ 33), respectively. During surveys, many complaints were noted about unfair distribution of those amounts. Many households reported that they had to provide bribes to UP members and their cadres in order to receive that money. One household-head claimed that “sometimes the UP Members kept half of the total allocated amount”. Another household-head reported that “my house was destroyed, but I didn’t receive any cash, because I couldn’t provide any bribe to the UP Member”. One local key informant claimed that he gave Tk. 5000 as a bribe to a UP Member for a new house (locally known as Saudi House—a donation from the Saudi Government, approximate cost of Tk. 25,000, US$ 325), but did not receive the house. He finally got the money back; however, many people did not. Similarly some affected households had to provide bribes to receive building materials. For example, one household-head said that “providing a bribe, I received two bundles of tin (corrugated iron sheet) because my house was destroyed”. Asian Development Bank (2004) found that the resource distribution through UP is less transparent, elite (UP representatives, and local political leaders) biased, and corrupt, which our findings confirm.

Significant numbers of household-heads identified many village headmen and village political leaders, who work as the local political agents of UP representatives, as corrupt (Table 2). UP representatives are highly dependent on the village headmen and village political leaders for votes during the election, because the local political leaders are the key individuals that influence voters. Consequently, UP representatives do not prepare relief distribution lists alone—they usually select the affected households together with the village headmen and village political leaders. As a result, household-heads keep close to the village headmen and village political leaders. Village leaders sometimes take advantage of this influence and demand bribes for putting local peoples’ names on the distribution list, which disrupts the fair distribution of recovery support. One household-head commented “the corrupted persons are the village leaders, who usually prepare the relief distribution list of the affected households”.

Another expression of corruption is substitution of lower quality goods for higher quality goods provided by international relief organisations, and distribution of only a proportion of relief goods received. For example, some household-heads reported that out of every 30 kg of VGF rice, 20 kg were rotten. What sometimes occurs is that parties responsible for relief distribution purchase supplies for distribution locally, whilst keeping the rice (or other goods) provided by relief agencies. Another household-head said that “UP distributed only 10–15 kg rice per household, while the allocated amount was 30 kg”. Another variation is irregularities in the length of relief support. FGDs reported that one VGF cardholder household is supposed to receive support for up to 2 years, however, after receiving bribes, some UP Members extended some beneficiaries’ support illegally. Relatedly, sometimes UP members recruited new supporters by giving benefits intended for vulnerable households to new households. Local journalists also confirmed these conclusions about relief irregularities.

Although the presence of corruption is widely recognised, a few disaster practitioners suggest that it is a significantly smaller problem than our analysis has concluded. For example, one disaster practitioner said that people sometimes exaggerate corruption. He claimed that UP-level corruption is at a minimum. He also added that if we distribute all relief goods through UP with strong monitoring, there might be a negligible level of corruption. He reported “we worked through a UP after Sidr and didn’t see any irregularities, due to strong monitoring”. However, UP representatives are aware of corruption, and aware of its influence on voting—in particular, not being corrupt can be difficult, because some villagers commit their votes based on whether they are able, via bribes, to get preferential access to relief goods and cash. Notwithstanding the dissenting comments, many lines of evidence—from each of FGDs, household surveys, local KIIs, and local journalists—confirm that UP representatives are involved in post-disaster corruption.

The volume of corruption after disasters is higher than in normal periods, simply because there are large flows of cash and relief goods into the affected areas. Similarly, Mahmud and Prowse (2012) found that post-disaster relief interventions are particularly prone to corruption because the flow of resources is rapid and substantial. These large money and relief goods flows encourage the UP and village leaders to act corruptly. However, this creates mistrust between the local people and UP leaders, i.e. these key linking relationships are damaged. From a disaster recovery perspective, this distrust is very concerning, because collective decision-making is important during the recovery process. Communities with more trust and stronger networks with local authorities can better bounce back after a crisis than fragmented and isolated communities (Chamlee-Wright 2010).

3.4 Role of UP Union Disaster Management Committees (UDMC)

The UDMCFootnote 3 is formally responsible for local-level disaster management. UDMCs are set-up by Union Parishad, as part of national arrangements for disaster management. The Standing Orders on Disaster (SOD—issued by the national government) give an active coordination role to the UDMC—it states that as part of its disaster responsibilities, the UDMC shall meet once a month during normal times; and that at the warning and pre-disaster phase, each committee shall meet more than once a week; and that during the disaster, shall meet as and when needed, even daily. In the recovery phase, committees are directed to meet once a week (Disaster Management Bureau 2010).

In our study areas, the UDMC was not carrying out that coordination role effectively. One of UDMC’s mandated roles is to provide pre-disaster awareness programmes, and another is to coordinate relief works (Disaster Management Bureau 2010). Yet in our study area, 75 % of household-heads did not know that the UDMC existed as a Union Parishad body (household survey). Most of the household-heads had not seen or heard of UDMC activity in their area (household survey, FGDs). Our interviews with disaster practitioners, and other research, indicate that this is the usual case in Bangladesh. During normal times, many UDMCs meet less often than once every 2 months, and some meet less often than once every 6 months; some also make no contribution at all to disaster preparedness (DPIs). In a study of similar region, Patuakhali, it was found that the UDMC made little contribution to disaster response coordination, and that most engagement by UDMC members was motivated by the opportunities to influence distribution of relief favourably for themselves and their networks (Islam 2010). In a recent study, Khan (2013) similarly found performance of the UDMC to be “weak” at all phases of disaster management, i.e. prevention, preparedness, response, and recovery.

4 Factors contributing to UP’s weak performance

4.1 Political factors

Political factors played a key role in UP’s poor performance. Fifty-five percent of household-heads identified abuse of political power and exchange of bribes as causes of UP’s inappropriate performance (Table 3), and 33 % of household-heads thought that conflict among the villagers was a cause of UP’s weak performance (Table 3).

Table 3 Responsible factors for weak performance of UP.
Full size table

The political needs of leaders played a central role in the inequities and corruption in disaster relief. Politicians depend upon villagers’ votes, and so many tend to distribute emergency relief and other recovery support in ways that reward their political supporters. One disaster practitioner commented, for example, that “UP representatives have political accountability at the local level. They can’t make the relief distribution list very objectively always, as their political priority is voters.” FGDs, DPIs, and local KIIs described how UP representatives were, at times, unable to agree on how relief goods should be distributed, with the outcome being that supporters of the political majority were favoured. A second consideration for some politicians is the financial strain of electioneering. This study encountered an “open secret”: Some UP representatives had spent money as an “investment” before an election to gain political power and used relief and rebuilding activities as an opportunity to recover that money and also to save for the next election (DPIs, local KIIs).

From a social capital perspective, one way to understand this favouritism and corruption is as politicians taking steps to build their social capital (through favouring partymen and buying voters)—but a form of social capital—“uncivil social capital” (Callahan 2005; Pérez-Díaz 2002)—that tends to undermine the development of broader “civil society” (Putnam 1995). For the purposes of this study, with its focus on relatively local social capital in Bangladesh—at the resolution of a Union Parishad—we differentiate “civil” from “uncivil” social capital on the basis of whether or not the form of social capital being built supports development of civil society (through social connectedness and civic engagement based on equity and trust) (Putnam 1995) at the Union Parishad level. From this perspective, politicians building networks of support through favouritism, and through taking bribes for relief provision (n.b. that there is certain mutuality in this, from many households’ perspectives), are building “uncivil social capital”: social capital that supports them and their party and undermines the community as a whole. Relief distribution that is managed with integrity and equity, on the other hand, would be building “civil social capital”. We identified instances of civilly building social capital (in this sense) through relief distribution in some other regions, but not in our study area (DPIs, FGDs). In these villages, at this time, patronage relationships were dominant.

The “incivility” of local dependence on patronage relationships is obvious at the local level. A consequence of political leaders favouring their own party’s people (which happened much more during recovery support—during the emergency relief phase when people were very vulnerable, the consequences of favouritism are higher, so inequity was lower), was conflict amongst villagers (FGDs). This conflict at village level, between people of different political ideologies, further hindered the equal distribution of recovery support (DPIs, local KIIs). When “relatively local” social capital is built in a way that undermines broader civil society, it tends to slow down economic growth and social development (Svendsen 2003). Putnam (2000) refers to this as the “dark side” of social capital, which promotes “unsocial capital” through doing “public bads” rather than “public goods” (Levi 1996).

These weaknesses are expressed institutionally. For example, UDMCs’ internal organisation is weak, and their activities highly politicised. A local key informant said that “the UDMC members don’t have any power. The Chairman is all in all, his decision is final, and the other members work at the will of the Chairman”. A disaster practitioner commented, resonantly, that “most of the UDMC members don’t know about their membership. Often the Chairman decides who is a member of the UDMC, chooses his party supporters, doesn’t involve them in UDMC activities, and simply makes the UDMC’s decisions himself”. Speaking about UDMCs in general, another disaster practitioner said that UP’s major recovery activities are run through the Chairmen and their partymen; involvement of formal UDMC members is limited.

Finally, Cyclone Sidr occurred during a period in which a Caretaker Government, supported by the military, was ruling Bangladesh. During this period, many national and local political leaders, including UP representatives, were sent to jail, or fled, after being accused of corruption (often legitimately, and sometimes as political harassment). Consequently, some UP representatives were simply unable to contribute to disaster relief and recovery efforts.

4.2 Lack of financial and logistic support

Weak financial capacity is also a prime factor contributing to UP’s poor performance. There is no national government budget allocation to UP to enable it to maintain its regular activities, and it is also difficult to collect tax from poor villagers (Meeting with UP staff). Problems with revenue collection undermine UP’s capacity to provide services of all kinds (Asian Development Bank 2004). An example of this, which illustrates the problems this causes for disaster resilience, is issues with the salaries of chowkidars who provide crucial early warning services, and run disaster preparedness programmes. Chowkidars are notionally employed full time; however, their monthly salary is only Tk. 1900 (US$ 27), about one-fourth of average income (about Tk. 8000) in the villages (household survey). Moreover, payment of their salaries is not reliable, due to the income difficulties experienced by UP. As a result, they undertake considerable casual work in order to support their families. This undermines the quality of their work.

The financial difficulties also impact the chowkidars’ operational activities. The chowkidars play a vital role in disseminating early warnings, search and rescue, and distributing relief goods. However, they do not have the required logistic support, for example, megaphones (for disseminating early warnings), bicycles and mobile phones (for communication with the community), umbrellas and raincoats (for working within rough weather), torches (for working at night), and heavy knives, axes, and saws (for cleaning debris and uprooted trees after cyclones) (Meeting with UP staff).

The ineffectiveness of the UDMC also has some financial roots: despite its supposed centrality in coordination of recovery efforts, it is a voluntary body—UDMC members are not paid, and financial support for UDMC activities is poor. Another constraint is that the nationally funded Annual Development Programme grants, which fund reconstruction and rehabilitation, are very small at the UP level, compared to the grants provided to higher levels of local and regional government, for example, Zila (district) and Upazila Parishad, and the allocation of these funds is not always transparent and is prone to political manipulation (ADB 2004).

Broader issues of trust also impact UP financial capacity. Only national government resources are channelled through local government, while the bulk of resources (both from national and international donor organisations) are distributed through NGOs. In general, donors have higher trust in NGOs than in local government. This reflects the problems of corruption discussed above and is a further illustration of how weaknesses in “civil social capital” (Callahan 2005) reduce the capacity of UP to support disaster recovery.

4.3 Uncertainties about relief support

Local people faced considerable uncertainty about their future access to relief. Generally, they had little information about existing and future relief packages, about how long they will continue to receive support, and whether recovery support will be distributed fairly. Moreover, relief packages were not shaped on the basis of assessments of the community’s needs. For example, people’s needs were temporary shelters, but UP provided food, and sometimes households received raw food, yet had no cooking items like fuel or a stove to cook them with. Also, in UP’s relief distribution, the specific needs of the aged, disabled and children were not considered, and their needs were not prioritised [The NGOs, however, did prioritise these vulnerable groups (FGDs, household survey)].

In response to these uncertainties, coastal households sometimes followed unfair strategies to receive more relief goods. Focus groups and local key informants reported that when cropping seeds became available, some people presented themselves as farmers. These same people denied having any farming land during distribution or other relief support, as the landless got priority for relief. One disaster practitioner said, perhaps harshly, that “a class of looter, broker, and relief taker has emerged after Sidr, which led to bribery”. Another way to look at this is as an alternative survival strategy of the Sidr-affected poor people.

Local people’s vulnerability and uncertainty is also partly responsible for UP’s corruption. Supporting the exchange of bribes, one household-head said that “who can’t eat (take bribes), can’t bring (can’t broker access)” (je khaite parena, se anteo parena). Resonantly, one disaster practitioner commented that people sometimes want to elect a person who has record of corruption, as they think the person who can take bribes from the local people can also bring recovery projects through providing bribes to the central level. Another disaster practitioner said that sometimes, some people do not see corruption as negative; rather, they see bribes as usual practice. They believe that the UP representatives work hard to bring relief goods from the higher authorities. Consequently, answering a question on “giving a bribe to a UP Member”, a household-head asked a reverse question: “Member has given me relief. I should also give him something, shouldn’t I?” We consider that these people are not directly responsible for corruption; rather, social perception, local culture, and weak governance promote corruption. However, people’s perceptions of normal leadership at the local government level play a key role in maintaining these “uncivil” norms.

4.4 Lack of collaboration between UP and NGOs

Lack of collaboration between NGOs and UP is a further cause of UPs’ weak performance. Generally, NGOs are reluctant to involve UPs with their post-disaster activities. There are a number of reasons for this. The post-disaster relief and recovery capacity of UPs is poorer than that of NGOs, because NGOs receive more donor support than UPs. Local NGOs also have a negative attitude to UP and its allied bodies. Disaster practitioners said that in most cases, NGOs do not consider partnering with UP due to fear of losing autonomy. One disaster practitioner said that the national and international NGOs do not involve UDMCs, except in some consultation meetings, because they consider them to be ineffective and ornamental political bodies. Consequently, after a disaster, UP is transformed into a recipient of NGO’s support, not a partner.

5 Approaches to reforming local government practice

Reducing corruption in UP’s contributions to recovery is the central challenge that this analysis presents. Lower levels of corruption offer the potential of fairer—more strongly needs based—distribution of relief and support for recovery of livelihoods. Moreover, from a legal and national policy perspective, UP should, in principle, be the central actor in implementing all disaster risk reduction and recovery activities at the local level. There are also concerns that NGOs’ level of involvement tails off rapidly after disaster recovery funds are expended, whereas UP remains better connected to the community, as UP is a democratic organisation and UP representatives are directly elected by the voters. One disaster practitioner commented, for instance, that the NGOs return home after completing the funeral activities (chollisha—the fortieth day) after a disaster.

There are formidable challenges to reforming local government practice; however, as corruption is entrenched in political life in Bangladesh. In a survey of perceptions of corruption in Bangladesh, Hardoon (2011) found that 66 % of respondents reported that they had paid a bribe in the previous 12 months, with police and the judiciary being the most frequent recipients of these bribes. This was substantially higher than other South Asian countries. Moreover, general reform of UP in Bangladesh is necessarily integrated into reform of the state, because as (Tanzi 1998, p. 588, c.f.; Zafarullah and Siddiquee 2001) emphasises, “some of the measures to reduce corruption are at the same time measures that change the character of the state”. This in turn is typically seen as requiring strong, uncorrupt political leadership. Tanzi (1998, pp. 590–591) asserts, for instance, “any serious strategy to attempt to reduce corruption will need action on at least four fronts [… first and foremost] honest and visible commitment by the leadership to the fight against corruption, for which the leadership must show zero tolerance”. Resonantly, speaking of Bangladesh specifically, Zaman et al. (2011, p. 11) conclude, “Political commitment is the key, supported by the capacity to enforce the law without fear or favour. It is often argued that strong political commitment is an essential precondition for the effective implementation of the UNCAC [United Nations Convention against Corruption]. Political consensus across the political spectrum in general, and in the two major parties in particular, is also mandatory for better implementation of laws”.

From the perspective of social change, this advice is somewhat paradoxical: political leaders for whom acting corruptly has become normalised are to become leaders who drive anti-corruption measures with integrity. Bangladesh has legislation and institutions in place to reduce corruption; the problem lies with their implementation (Zafarullah and Siddiquee 2001; Zaman et al. 2011). Transparency International Bangladesh lobbied for Bangladesh to sign on to the UNCAC, which the government did in 2007 (Zaman et al. 2011). This increases external scrutiny, which is a useful contribution to reform. However, many changes implemented in parallel are generally considered to be necessary for wholesale reform in circumstances like those in which Bangladesh finds itself (Tanzi 1998; Zaman et al. 2011), and what Bangladesh needs is a political willingness to use its anti-corruption laws and institutions energetically and fairly. That willingness exists at the grass roots level. Hardoon (2011, p. 18) found, for instance, that in Bangladesh over 80 % of respondents said that they could “imagine [themselves] getting involved in fighting corruption”, “would report an incident of corruption”, and “would support their colleague or friend, if they fought against corruption”. More than 90 % of household-heads we interviewed stressed the importance of arranging an awareness campaign through discussion meetings on corruption in recovery activities (household survey), and this emphasis found general support (FGDs, local KIIs, DPIs, and PMIs). There is, therefore, a platform in community sentiment for a changed approach to corruption—notwithstanding that some accept it as normal. Perhaps surprisingly, in our household survey, only 20 % of household-heads were not hopeful of controlling corruption, because they saw the corruption-controlling local law-enforcing authorities (e.g. police, and Upazila administration) as also corrupt.

In the absence of broader reforms, there are still some opportunities to pursue reduction in corruption in UP’s disaster relief and recovery practice. We identified effective local action to reduce corruption and improve performance in NGOs’ operations (Islam and Walkerden 2014). Experience elsewhere indicates that sometimes strong leadership can markedly reduce corruption within a particular arena (Bardhan 1997). At the UP and village levels, there are civil society leaders who may be able to push for reduction of corruption—our interviewees (household survey, FGDs, DPIs) all identified local Imams and teachers, and women’s representatives, as good candidates, and identified NGOs, FBOs (Faith-Based Organisations), and other CBOs as substantially less corrupt than UP.

Directions for improvement at the local level include:

  • About 85 % of household-heads emphasised the need to increase coordination between UP and other local actors, e.g. local NGOs and CBOs, and Imams (as FBOs’ representatives), village headmen, and school and madrasa teachers (household survey).

  • About 95 % of household-heads said that the UP representatives should consult with the local people to identify community needs before and after disasters (household survey)—in effect they are asking for a participatory development approach (Chambers 1983) to relief distribution. [To identify actual vulnerable households, UP, working with other local leaders, could consider the fitra Footnote 4 receivers of a certain village, as the poorest people usually receive fitra and people generally do not provide wrong information about fitra receivers.]

In the absence of locally, regionally, or central driven reform of government, it is probably best to follow de facto current practice and rely primarily on NGOs for relief distribution and provision of support for recovery of livelihoods (c.f. the recommendations made in Islam and Walkerden 2014). Strengthening NGOs’ capacity is less challenging than strengthening local governments’, and local people can opt to work with local NGOs that have more integrity, in a way that they cannot do with local government.

6 Conclusion

On the Bangladeshi coast, Union Parishad bears significant responsibilities for disaster recovery, including roles for immediate response through to relief distribution, shelter and livelihood assistance, emergency information, reconstruction of major community services, and so on. In practice, patronage relationships (notably favouring political supporters) and bribery play a substantial role in how those responsibilities are discharged. Nonetheless, many households received emergency relief and help with rebuilding from UP (Table 1). However, the equity and efficiency of these contributions to recovery were markedly diminished by corruption.

In principle, reducing corruption in UP’s contributions to disaster recovery could significantly improve resilience. The difficulty, however, is that corruption is widespread in government administration in Bangladesh. To improve it in UP generally, and thence in disaster resilience and recovery work specifically, widespread reform is needed. Community sentiment would support that. It is not a change that cyclone risk would drive; however, disaster recovery would be one beneficiary of a general shift.

There are possibilities for exemplary local action in recovery, led by (for example) Imams and women’s representatives. However, in general it appears that continuing to look primarily to NGOs to distribute relief and recovery support is appropriate. Villagers can express their preference for lower corruption by changing the NGOs they work with, and reforming and strengthening NGO practice is easier (Islam and Walkerden 2014).

Although the current study is based on two study villages, the findings regarding UP’s contributions to disaster resilience and recovery are expected to apply generally along the Bangladeshi coast, as the hazards and vulnerabilities, the villages’ socio-economic characteristics, and the capacities of UP are similar throughout this region (Alam and Rahman 2014).

The empirical findings of this study provide a new understanding of how UP, as a linking social network actor, works at the post-disaster rebuilding stage and in fostering resilience. This research adds to the growing body of the literature on the contribution of local government to disaster resilience and recovery. Additional research on how corruption might be reduced, in disaster recovery activities and generally—e.g. by drawing lessons from locally, regionally, or sectorally successful efforts in Bangladesh and elsewhere—could extend this work in important ways.