reading review

lable at ScienceDirect

Habitat International 52 (2016) 57e66

Contents lists avai

Habitat International

journal homepage: www.elsevier .com/locate/habitat int

Disaster governance and the scalar politics of incomplete decentralization: Fragmented and contested responses to the 2011 floods in Central Thailand

Danny Marks a, *, Louis Lebel b

a School of Geosciences, The University of Sydney, NSW, 2006, Australia b Unit for Social and Environmental Research (USER), Faculty of Social Sciences, Chiang Mai University, 239 Huay Kaew Rd, Chiang Mai, 50200, Thailand

a r t i c l e i n f o

Article history: Received 11 August 2015 Accepted 14 August 2015 Available online 1 October 2015

Keywords: 2011 Thailand floods Disaster governance Politics of scale Incomplete decentralization Thai flood management

* Corresponding author. E-mail address: [email protected] (D. M

http://dx.doi.org/10.1016/j.habitatint.2015.08.024 0197-3975/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

This paper applies the politics of scale as a framework to examine how decentralization reforms and the associated power relations between government agencies at different levels affected disaster risk out- comes in Thailand, particularly during the 2011 floods in Central Thailand. It argues that Thailand's decentralization has been incomplete due to the retention of power and resources by central bureaucrats and the continued weak capacity of local administrative organizations (LAOs). In addition, the country's overall fragmented and politically polarized governance has hindered policy coherence at all levels, including the local level. Incomplete decentralization alongside persistent fragmentation along minis- terial and sectoral lines has undermined disaster governance and distributed risks unevenly and unfairly. The governance weaknesses visibly materialized during the 2011 floods. Except for the Bangkok Metropolitan Administration (BMA), LAOs once again had insufficient capacity to effectively respond to the floods and were given insufficient assistance by the central government. The central government sought to monopolize power, did not consult local communities, had limited capacity to enforce all of its decision, distributed risks unevenly, and overall performed poorly. Similarly, the BMA dominated other much smaller local government units within and beyond its formal boundaries.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

A common view in international discussions is that there is a need to further decentralize disaster risk reduction (DRR). The Hyogo Framework for Action encourages governments to “recog- nize the importance and specificity of local risk patterns and trends, [and] decentralize responsibilities and resources for disaster risk reduction to relevant sub-national or local authorities, as appropriate” (UNISDR, 2005, 6). The expected advantages of decentralization, in general, include: improved efficiency, as local governments can be more responsive to needs of local residents; more equitable access; and improved downward accountability, political engagement, and community participation (Faguet, 2014). The latter requires decentralization, as we define it, to be not just an administrative or fiscal redistribution to local levels, but also a devolution of political decision-making power (c.f. Litvack, Seddon, & Ahmad, 1999).

arks).

Greater local participation in decisions is expected to lead to more appropriate and sustainable DRR interventions (Wisner, Blaikie, Cannon, & Davis, 2004). A multi-country comparison of disaster losses found that those with more decentralized governments had fewer fatalities (Skidmore & Toya, 2013).

Empirical research, however, suggests that the expected bene- fits of decentralization for DRR, especially in the case of flooding, are not always forthcoming. There are several reasons. First, countries often devolve responsibilities but not skills or human and financial resources. Flood disaster legislation in Vietnam, for instance, acknowledges the importance of local communes in disaster responses and DRR, but an analysis of its implementation found a lack of training, and of technical and financial support, from the central level to the local levels (Chau, Holland,& Cassells, 2014). In Indonesia, decentralization reforms have created opportunities for local authorities to play a greater role in DRR but both local-level capacities and commitment to integrate DRR into development remain major constraints (Djalante, Thomalla, Sinapoy, & Carnegie, 2012). While the framework of a multi-level institutional archi- tecture is now formally in place for disaster preparedness, including

D. Marks, L. Lebel / Habitat International 52 (2016) 57e6658

early warning systems for tsunamis, implementation remains a challenge (Chang Seng, 2013) and there is still limited integration with coastal development laws (Thomalla & Larsen, 2010).

Second, local traps can arise. An overly narrow focus on local perceptions, knowledge and interests may become a barrier to recognizing solutions from other locations or available at other levels (Brown & Purcell, 2005). For example, in Greater London decentralization and increased public consultationdby empha- sizing more strongly the views of residents directly affected by floodsdled to increased support for conventional structural pro- tection measures, and less consideration of alternative adaptation options and the needs of the wider population (Harries & Penning- Rowsell, 2011). In terms of risks, public opinion pushes for reducing the likelihood of events rather than reducing vulnerabilities or the magnitude of impacts. Likewise, a study of social learning processes around the flood-prone municipalities in Kristianstad, Sweden, concluded that decentralization limited learning at the national level (Johannessen & Hahn, 2013).

Third, decentralization does not mean that other levels, and coordination between them, are no longer important. Consider two examples. Established after major floods in the Rhone River, France, a new partnership brought together local and state authorities to tackle floodplain management at the river level (Guerrin, Bouleau, & Grelot, 2014). However, the initiative was abandoned after five years of negotiation, as a consequence of historical and unresolved politics of scale; different actors were unable to overcome differ- ences in interests and framings of the flood problem. Efforts to introduce river basin management in Mongolia were affected by national-level resistance to decentralization, delays in the wider political process, and politics of scale around delineation of basin boundaries (Houdret, Dombrowsky, & Horlemann, 2014).

These examples suggest that decentralization reforms often struggle to deal with coordination and political issues related to scale. Actors at different levels do not want to be coordinated by someone elsedespecially if they are a department in another ministry, and thus part of another hierarchy. As a number of scholars have argued, the most effective and equitable form of water governance is polycentric or “decentralization with coordi- nation” (Pahl-Wostl, Lebel, Knieper, & Nikitina, 2012). For a response to be polycentric, groups must have the ability to solve problems on their own that are institutionally enabled in a self- governance system. Polycentricity also denotes many decision- making centers independent of each other but that can make mutual adjustments.

The politics of scale refers to all the different ways actors contest scale choices. Scalar arguments often fit with organizational in- terests, and the levels in which power are wielded, resources ac- quired, and to which blame and unwanted responsibilities are shifted (Lebel, Garden, & Imamura, 2005). Arguments about scale are also often about boundaries: what should count as ‘in’ or ‘out’. As Brown and Purcell (2005, 608) argue, “scalar configurations are not an independent variable that can cause outcomes, rather they are a strategy used by political groups to pursue a particular agenda.”

Scalar politics are thus both a cause and consequence of decentralization. Decentralization is often a response to pressures from society to “re-level” political and bureaucratic hierarchies. These bottom-up pressures on local governments, especially when they include demands from vulnerable and lower-income groups, are important in building resilience to multiple risks including climate change (Satterthwaite, 2013). Once some decentralization reforms are undertaken, this often creates pushback over the ‘right’ level for decisions, money allocations, and accountability, and about who should be responsible for what and who should do the work. This resistance is one reason why decentralization is often “incomplete”, and why efforts to decentralize disaster management

can fail to increase local agency in practice (Blackburn, 2014). Apart from the politics of scale, when it comes to management

of floods, there is also a politics of position (upstream versus downstream) and place (Lebel et al., 2005). Capitals, it is often argued, as economic and political centers, need to be protected in the ‘national interest’; in practice, this means that risks and flood waters must be unevenly distributed to other areas (Lebel & Sinh, 2009). This example is also an instance of re-scaling, underlining how spatial politics has multiple dimensions (G€org, 2007), and that space and network hierarchies may be distinct and interact with each other (Bulkeley, 2005).

This paper analyzes the consequences of incomplete decen- tralization reforms in Thailand for the governance of flood-related disaster risks, with a special emphasis on the role of the politics of scale in the responses to the 2011 floods in Central Thailand. Three specific questions, with an emphasis on factors and processes relevant to the management of flood disasters, are addressed: (1) How has decentralization proceeded in the last two decades? (2) What are the implications of this history for disaster risk man- agement? (3) What role did the politics of scale and decentraliza- tion have in the responses to the 2011 floods? The paper uses a mixture of primary and secondary sources, drawing from in- terviews conducted with government officials, academics, and community leaders from March 2014 to January 2015.

2. Decentralization and scalar politics in Thailand

Until the 1970s, Thailand's political system could be described as a bureaucratic polity: bureaucrats and the military dominated the political process and monopolized power at the local level through powerful central government ministries, particularly the Ministry of the Interior (MOI). Local administrative organizations (LAOs) were highly centralized, since MOI-appointed provincial governors controlled most of the appointments of local officials.

In the 1970s, three changes occurred that led to the decline of power of the bureaucratic polity and the rise of local politicians- cum-businessmen power (Shatkin, 2004). First, after middle class Thais agitated for change and launched large-scale street protests, the national government increased the budgets of the elected Provincial Administration Organizations (PAO), and established the Bangkok Metropolitan Administration (BMA). Second, democrati- zation occurred from this period onwards (until 2014) as the power of parliament grew. Third, industrial transformation in the prov- inces provided new opportunities for provincial elites to accumu- late wealth and translate this wealth into political power (Shatkin, 2004), and also created growing popular demand for local partici- pation and empowerment.

The increased influence of local politicians, combined with democratization, paved the way for Thailand's decentralization reforms in the mid-1990s. Politicians, together with the public, civil society organizations and prominent academics, advocated for decentralization as a way to improve transparency and account- ability. These forces for reform were successful in enshrining decentralization in the 1997 Constitution, and in pushing the pas- sage of the Decentralization Act in 1999 (Dufhues, Theesfeld, & Buchenrieder, 2014). However, as the following subsections show, decentralization has not progressed smoothly, and remains incomplete in terms of the transfer of administrative, fiscal, and political responsibility from the central government to subordinate governments.

2.1. Administrative decentralization

Similar to the French system, the 75 provincial governments in Thailand are currently organized under two different, parallel

D. Marks, L. Lebel / Habitat International 52 (2016) 57e66 59

structures: a centralized provincial administration, headed by a governor who is appointed by the Ministry of Interior; and an autonomous provincial administrative organization (PAO), whose head is elected. Under the PAO are autonomous local administra- tions (LAOs): municipalities and rural sub-district administrative organizations (SAOs) (see “Fig. 1” below). The heads of LAOs are elected. However, while the number of LAOs have proliferated, the central government has granted them very little independence and administrative responsibilities. Provincial governors and district chief officers must grant permission to LAOs before they can pass local laws, budgets, and development plans. Central government agencies have kept many responsibilities, and LAOs are unable to provide public services unless they are mandated (Nagai & Kagoya, 2008).

Further, governors and district officers who are part of the central government apparatus still oversee the provincial and lower-level autonomous offices. As a mayor in a municipality in Pathum Thani Province said, “The municipality is under the pro- vincial government. I need to wait for their orders to make a de- cision.” He also complained that he had to coordinate with many agencies to provide public services, which slowed down the de- livery drastically. For example, to dredge a canal, permission had to be sought from the Royal Irrigation Department (RID). He declared, “My office should have our own authority to manage these kinds of public utilities by ourselves” (personal communication, January 5, 2015). Moreover, although progress was made in transferring re- sponsibilities for services to local governments, less attention was given to transferring the corresponding human resources to local authorities (Nagai & Kagoya, 2008).

2.2. Fiscal decentralization

Fiscal decentralization has also been incomplete. Although the Decentralization Act decreed that in 2006 LAOs' revenue should be 35% of total government expenditure, the actual percentage in 2011 remained at around 26% (Kokpol, 2011). Moreover, local govern- ments have been granted very little authority to raise their own revenue, being able to collect only three types of smaller taxes, which in 2006 consisted of, on average, only 9% of their total rev- enue (Nagai, Funatsu, & Kagoya, 2008). Many local authorities in Thailand have expressed dismay at this financial situation, lam- basting the national government for giving them responsibilities for additional services, such as infrastructure provision, planning, and natural resource management, without sufficient finances. A survey in 2008 found that 94% of LAOs felt that funding from the central government was insufficient (Nagai et al., 2008). They are also exasperated that “the local government still needs to ask for

Fig. 1. Forms and configuration of LAOs in a Thai province (Department of Local Administration, 2010).

approval from the provincial governor and if he cannot make a decision, he has to ask the central government to approve the request” for funds. (PAO head, personal communication, January 9 2015).

However, a related problem is that the majority of LAOs are too small to become financially viable and have insufficient capacity to provide services. For example, in 2009, Ban Chiang SAO spent less than a dollar per person on health care and environmental pro- tection (World Bank, 2012). Another problem is that top-down and bottom-up accountability are hampered because information on service delivery and fiscal performance is not made publicly available. In 2010, the MOI's financial reporting system only had fiscal information available for about 1200 of the 7854 LAOs (World Bank, 2012). Consequently, it was difficult to measure their accountability and the quality of financial management.

Overall, this financial system has further weakened the auton- omy and capacity of LAOs. Not only has their limited authority to collect money created a high level of local dependency on patronage ties with national government bureaucrats and politi- cians, it also means LAOs do not have large enough budgets to fulfill their responsibilities.

2.3. Political decentralization

The heads of all local administrations were elected from 2003 (with a brief hiatus from 2006 to 2007) until the 2014 coup, when elections were suspended indefinitely. Political decentralization has also been limited, flawed, and often captured by the elite.While the dual administrative structure was created with the objective that elected LAOs would represent constituent interests, and centrally-appointed officials would effectively manage LAOs, this aim is not being met. Rather, while the heads of LAOs are still held accountable by their constituents, most decision-making power lies with unelected central-level officials. This has blurred account- ability and has created tension between the central and local gov- ernments. When disagreements arise, the decisions of central officials usually trump those of LAOs (World Bank, 2012).

Another trend is that local elites have taken advantage of the opportunities created by political decentralization to increase their power and wealth. For example, in Buriram province, the Chidchod family further consolidated their power within the province. Pre- viously members of their family had been MPs and Senators in the province for decades and, after the country decentralized, they not only remained MPs but also their network members became the PAO head and won 86% of the PAO council member seats in 2003. This monopolization allowed them to access LAO budgets and grants from the national government. In 2005, the public in Bur- iram filed over 60 complaints about conflicts of interest in LAO construction projects (Wongpreedee, 2007).

Decentralization in Thailand has so far neither resulted in any major devolution of power to local communities, nor led to any fundamental changes in central-local power relations (Dufhues et al., 2014; Sato, 2011). After a strong pull towards decentraliza- tion, central bureaucrats, particularly provincial governors, through different forms of pressure, influence and financial tools, have resisted and distorted the implementation of these reforms and retained a high degree of control at the local level. Concurrently, local elites have consolidated their power in many areas. Decen- tralization reform in Thailand, therefore, should be seen as a scalar struggle between actors seeking to promote their own interests. Especially after the 2014 coup and the subsequent dismantling of some local committees and the suspension of local elections by the military government, the central government has continued to concentrate power at the central level. However, fragmentation and lack of collaboration arise not only between the central-local levels,

D. Marks, L. Lebel / Habitat International 52 (2016) 57e6660

but also at the central level. In this multi-centered autocracy or “an entangled mass of interlocking relationships, alliances, and strug- gles between and among many centers of power often in compe- tition with one another” (Reynolds, 2012, 2), each ministry and department operates as a small kingdom, trying to control the is- sues and projects under its jurisdiction. This fragmentation has caused interagency clashes over budget allocations and control of projects in which scale politics recur. The fragmented centralized system which has disaggregated power and hindered policy coherence makes it difficult for LAOs to effectively provide services and govern their jurisdictions.

3. Implications for disaster risk governance

The implications that incomplete and contested decentraliza- tion reforms, and the on-going politics of scale in a fragmented bureaucracy, have for disaster risk governance in Thailand are examined in this section. The analysis is organized with respect to the land and water sectors. The two sectors have been chosen because of their importance in exacerbating and reducing flood risk (Wheater & Evans, 2009).

3.1. Land sector

The country's form of decentralization has contributed to the lack of land use planning in Central Thailand (see Fig. 2 below) and made these areas more at risk to floods. Locally-elected leaders have often responded to the electorate's desire to give higher pri- ority to infrastructure development, particularly roads, and income-generating projects over long-term conservation (Sato, 2011). Urban planners at both the national and sub-national levels have rarely considered flooding as a factor in their plan- ning (Tangwisutijit, 2007).

Fig. 2. Map of Bangkok Metropolitan R

LAOs and line agencies also rarely have an incentive to restrain urban growth by designating green areas or other limitations. Consequently, land use plans are “in practice mapping exercises that capture changes in land use that have already occurred, rather than setting future land use limitations” (Friend et al., 2014, 13). Or, as in the case of Ayutthaya Province, the provincial Office of Public Works and Town and Country Planning drafted a comprehensive land use plan addressing urban overspill but did not have any legal mechanisms to enforce the plan (Maneepong & Webster, 2008). Thus, in most instances, land use plans have not been enforced. As a result, it is common to find different land uses next to each other in urban and peri-urban areas. Further, when they do implement land use plans, LAOs often do not coordinate the plans with each other, which can weaken their effectiveness. For example, BMA has designated a water catchment area to be left undeveloped in northwestern Bangkok, but north of this area, the Nonthaburi provincial government allowed extensive development, thereby rendering this catchment area ineffective (Marks, 2015).

The rise of the local political elite has also contributed to haphazard urbanization and increased vulnerability to flooding. First, large landholders and real estate developers have become elected or are allies of those who are elected. As a result, these local leaders often have not sought to regulate land codes and instead weakened their enforcement (Kritsanaphan & Sajor, 2011). Second, they have often either acted as brokers for outside investors wanting to buy a big piece of land, or invested in a number of housing and industrial projects. They would consolidate land for these projects either through legal purchases or graft and intimi- dation (Chantornvong, 2000). Their excessive pursuit of property development has contributed to overbuilding in Bangkok, partic- ularly in green zones (where such development is illegal), and in floodways, such as industrial estates in Ayutthaya. Four northeast districts of Bangkok were planned to be floodplains, but residential

egion (CartoGIS Coordinator 2015).

D. Marks, L. Lebel / Habitat International 52 (2016) 57e66 61

and industrial development converted these previously-agriculture areas into urban areas (Wijitbusaba, 2013). During the floods in 2011, water diverted to these green zones not only severely damaged the buildings therein, but furthermore, these buildings blocked the water from flowing south into the Gulf of Thailand. Third, local politicians have protected illegal communities encroaching upon canals in order to stay popular. According to a deputy governor of one of Bangkok's northernmost districts, this is why a number of communities are still allowed to encroach upon a major canal, Prem Prachakorn (Don Muang District Deputy Governor, personal communication, July 16 2014). Last, in public hearings and meetings between government officials and business interest groups, landowners and developers have pressured gov- ernment officials to change land-use plans from green zones to other areas so that they can build in these areas (BMA planning official, personal communication, April 4 2014). This pressure has contributed to the decline of green space in urban areas.

3.2. Water sector

Just as in the land sector, Thailand's incomplete decentralization and scalar politics have contributed to increasing risks of flooding, and unevenly distributed this risk. Similar to the country's overall administration, the water sector is fragmented, consisting of numerous agencies at the central level, and has weak local bodies. In the Chao Phraya River Basin, the government has created eight river basin communities (RBCs). While this arrangement has encouraged grass roots participation and inclusive and collective active solutions, the water management system is still highly centralized and RBCs have little power. They have sparse financial and human resources, and their roles and responsibilities are un- clear. Moreover, they lack a legal mandate: government agencies are not obligated to follow plans and proposals or include them in their budgets. Thus, RBCs have limited capacity themselves to implement projects and so far have been mostly unsuccessful. Further, at the national level, the Department of Water Resources (DWR) has inadequate power to fulfill its regulatory role and integrate river basin planning throughout the country (Asia Development Bank, 2012).

Moreover, none of the approximately 30 departments involved in water resource development have an unambiguous mandate to manage flooding. As Thai academic Nipon Poapongsakorn (2014, personal communication, December 29, 2014) asserts, these “de- partments carry out their own function and do not look at national interests. Departments are like a horsedthey look only straight ahead.” For example, RID is responsible for protecting irrigated areas from flooding, but is not responsible for land outside these areas. In some instances, it has allowed flooding in urban areas in order to protect agricultural ones (Manuta, Khrutmuang, Huaisai,& Lebel, 2006). Before 2011, local government public work agencies often constructed dykes in an ad hoc effort to address local flooding problems without using any integrated plan. However, since many dykes function as access roads, the Department of Rural Roads is responsible for their maintenance (Manuta et al., 2006). The con- struction of dykes along riverbanks in many upstream provinces has reduced the capacity of dykes downstream to handle a 100- year flood. Additionally, provincial or local governments have sought to reduce their risk of flooding by distributing it to other areas. For example, when BMA officials increased the height of a floodwall along the Chao Phraya River, they knew it would nega- tively impact neighboring Nonthaburi province. However, this area was outside of their jurisdiction (Lebel et al., 2009).

Scalar politics has also affected flooding in Central Thailand. The national-level RID and the provincial-level BMA have different objectives as was seen in 2011 and years beforehand. While BMA

seeks to protect its constituents by preventing any flooding in Bangkok by raising its embankments and closing water gates, these actions had made it difficult for RID to control flooding in upstream and surrounding provinces (Lebel et al. 2009; Molle, 2007). In another case, however, the national government trumped the provincial government. Although BMA had designated the wetland area around Suvarnabhumi airport as a green zone, the national government overrode this decision, built the airport there despite BMA's protests, and then protected it with its own polder-blocking canals. These canals have reduced the airport's risk of flooding, but heightened the risk for those living nearby, which is where the rainwater is pumped (Lebel et al. 2009).

4. 2011 floods in Central Thailand: government and community responses

In 2011 Central Thailand experienced some of the most severe flooding inmany decades. Overall in Thailand, the floods killed over 800, affected millions and cost the economy at least US$45 billion. Responses to the 2011 floods reflected the lack of polycentric governance and the uneven scalar relations. The country's elite prioritized keeping the inner city of Bangkok safe, as well as pro- tecting international and domestic capital, such as industrial es- tates. However, local politicians and community leaders contested certain elite decisions and the scalar configuration of the flooding by protesting, attempting to destroy sandbag walls or pry open water gates, or by protecting their own areas.

As a case study, the floods, which flowed from the north to the Gulf of Thailand, provide useful insights into how scalar politics affect disasters. First, it was a slow-onset event with unique polit- ical dynamics. Significant time existed for proposals and counter- proposals, actions and reactions, as well as deadlocks. For example, the city of Bangkok had a number of weeks to prepare, but some parts of the outskirts were still inundated for long periods while the inner city stayed dry. Second, the biophysical and ecological dimensions of the floods meant that actors could have a strong say on where to direct and retain the water. Since the Chao Phraya delta is flat and low, it is not difficult for actors to divert water, unlike had they been operating in steep terrain.

4.1. Lack of coordination

A number of examples show that the Thai state's response was not polycentric. First, as might be expected, LAOs lacked sufficient financial, technical and human resource capacity to effectively respond to the floods. A report by the Department of Disaster Prevention and Mitigation (DDPM) on the floods stated that the agency did not have adequate staff at the district and sub-district level, and “therefore, it took time to get the government personnel who belonged to other agencies to perform disaster- related activities at the field level” (Barber, 2012, 14). The report adds that because agencies did not have detailed guidelines to follow, their reaction was fragmented and depended on the orders of their superiors, some of whom lacked backgrounds in disaster management. Further, in many instances, DDPM provincial offices and LAOs did not trust each other (Barber, 2012).

Some local communities were upset at their LAO's response. For example, a community leader in Nonthaburi Province (personal communication, January 15 2015) complained that during the floods “the municipality was not helpfuldthey should have distributed food aid to peoplewho are in the areas that was difficult to reach. But they kept the food until it was bad. They also should have used boats to reach these people.” She further complained that community members received unclear information about the floods from the municipality: “It was still technical and hard to

D. Marks, L. Lebel / Habitat International 52 (2016) 57e6662

understand; the local government did not do anything to fit this information within the local context.”

However, someof the blamemust lie alsowith the provincial and national governments for not assisting LAOs sufficiently. As amayor inPathumThani (2015) said, “Themunicipality had tohelp itself; the national government didn't help much. They thought that the local governments would be able to handle the situation themselves.” A deputy clerk in a Nonthaburi municipality (personal communica- tion, December 12 2015) added, “None of the local governments were happy with the provincial governor. We had to use our own budget. They did not give enough pumps.” Overall, there was very little cross-level coordination or with adjoining LAOs.

Second, the central government seized administrative power and made decisions without consulting LAOs, which made it worse for many local communities. Under the Disaster Prevention and Mitigation Act, DDPM was given responsibility to ‘control and oversee disaster prevention and mitigation’ and in mid-August, DDPM set up emergency operations to respond to a ‘level three’ (or large-scale) disaster by setting up emergency operation centers at the national, provincial, and district levels. However, in early October, as the flooding worsened, Prime Minister Yingluck Shi- nawatra used the Civil Service Act to set up the ad hoc, highly centralized Flood Relief Operation Center (FROC), thereby margin- alizing DDPM and underutilizing its capacity (Barber, 2012).

FROC, headed by the Minister of Justice who had no background in flood management, made many mistakes. For example, it pro- claimed it could handle the floods, and that Bangkok would not be flooded. Although satellite imagery in late October showed that some parts of Bangkok faced severe, lengthy flooding, FROC incorrectly announced that the level of the water would decrease by the middle of November. Many residents, such as those in Don Muang, trusted the government's announcement and consequently did not protect their houses and possessions as much as theywould have done if had they been fully warned earlier (Marks, 2015).

FROC and other central government agencies also made mis- takes about where to release and block the water. For example, a mayor in Pathum Thani said, “Instead of using Rapeepat canal, which is higher than Rangsit canal, it used the Rangsit canal. But once the water came to the lower level, it was difficult to push it back. We had to sacrifice [this area as a flood zone] because of the government's decision.” FROC did not consult with any sub- national governments besides BMA about these decisions. This same mayor proclaimed, “FROC made the decision to block the water and told us that theywill do this, thoughwe told themwe did not agree. The process was not participatory.” He added that even the provincial governor did not have any say in the decision- making process after FROC was formed.

At the same time, FROC lacked the power or capacity to enforce its decisions. A strong example is that during the time Don Muang was heavily flooded, a FROC representative visited the Pathum Thani border of Khlong Ban Mai, where the municipality and local communities had blocked water from entering in order to protect themselves. The representative ordered them to open the nearby water gate to reduce the flooding in Don Muang, and said he would give them 500,000 Baht and two pumping machines. However, they refused and FROC backed down. A community leader in Don Muang was disappointed that FROC did not do more to push them to open the gate because he believed that Don Muang would not have been as heavily flooded if FROC had done so (Don Muang community leader, personal communication, August 28 2014). Another example is that in Bang Khen District in northern Bangkok, FROC wanted to open a closed water gate one meter immediately, but it took FROC ten days to negotiate with nearby communities to open it by 50 cm (DWR official, personal communication, October 7 2014). There were many similar instances where local politicians

and community leaders, seeking to protect their own turf, dis- obeyed the commands of FROC. They led their constituents to destroy flood protection dykes or open water gates so that water would be diverted to other areas. These actions undermined the overall response to the flooding.

Third, BMAwanted tokeep theareaunder its jurisdictionasdryas possible, and then, once some districts in the periphery of the city became flooded, to keep the inner city and its two industrial estates dry at all costs. According to RID and FROC officials, Bangkok could have taken in more water to lessen the burden on those flooded in northern and northwestern provinces. However, worried about the possibility of further rain, BMA did not want to take any more risks. For example, BMA opened water gates less than the level requested by FROC, or released only 10e15 cubic meters per second in some canals, when it had the capacity to release 50 cubic meters (FROC official, 2014). According to a member of FROC, the Prime Minister could have overruled the Bangkok governor Sukhumbhand but did not (FROC official, personal communication, July 18 2014). This conflict between BMA and FROC was shaded by the longstanding deep-seated polarization between the two main political parties: Pheu Thai of which Yingluck was the head, and the Democrat Party from which Sukhumbhand hailed. These leaders quarreled frequentlyandeach sought toblame theother for theheavyflooding.

Some district officers in Bangkok were not happy with BMA's decision to keep the inner city completely dry. One of them from Don Muang, a heavily flooded area, said, “As an officer I need to follow BMA's order, but as a personwho lives in DonMuang, I think BMA could have let more water flow through [inner] Bangkok” (Don Muang district officer, personal communication, June 30 2014). However, neither her office nor communities in Don Muang were consulted about this decision.

4.2. Urban communities fought back

Urban communities protested and contested scalar decisions made by FROC and BMA. While communities in Thailand are social and administrative constructs and a number have conflicts and divisions within them (see Elinoff, 2013), this section provides two case studies of inundated communities whose members became unified and worked together to contest the scalar configuration of the floods to reduce the level of water and shift the risk of the flood to other areas.

Yucharoen community, a lower-middle-income community in Don Muang, felt a deep sense of injustice about the placement of a wall of big sandbags that blocked water from leaving their com- munity. In Fig. 3, this community is located above the ‘big bag barrier.’

FROC did not consult with or inform affected communities that it was placing the big bag dyke there, nor did it tell them how long the bags would be there. After a month of living with smelly, waist- high water which had inundated all the houses and caused one elderly man to diedand finding out that the city's center remained drydcommunity members became angered and demanded that FROC remove the sandbag barrier. They began by protesting at the sandbag barrier. During one of the first protests, they held a public hearing and voted unanimously to remove the dyke. They also demanded that FROC respect their constitutional rights, including the right ofmovement (since the dyke blocked transportation along Viphavadi-Rangsit Road), and the right to hold a public hearing about the dyke. They then called the government to negotiate, and FROC's negotiator promised that FROC would remove the bags within a couple of days. However, when that day came, the gov- ernment had not sent anybody to remove the bags so the com- munities dismantled the dyke themselves. The level of the water fell rapidly once the bags were removed (Yucharoen community

Fig. 3. Map of placement of the big bag dyke in Don Muang. Yucharoen community is located along Viphavadi-Rangsit Road north of the dotted red line (Bangkok Post, 2011).

D. Marks, L. Lebel / Habitat International 52 (2016) 57e66 63

leader, personal communication, March 12 2014). Thus they con- tested the scalar configuration of the inner city remaining dry and the outer city being flooded.

Similarly, in Lum Luk Ka, communities were heavily flooded. In one slum community, Kukot Pattana, which is along the border of Lum Luk Ka district of Pathum Thani province and just outside Sai

Fig. 4. A map of the flood barrier along the BMA-Pathum Thani border. Ku

Mai district of Bangkok, residents felt that BMA's actions were unfair. According to one local resident, because BMA was fully protecting Bangkok, people in his area were ‘like floating ducks.’ After the community had been flooded for many months, 40 resi- dents, plus residents from other communities, protested at a big bag site (see “Fig. 4” below) that blocked the water from draining

kot Pattana is located north of the flood barrier (Frederickson, 2011).

D. Marks, L. Lebel / Habitat International 52 (2016) 57e6664

out of this area. They successfully persuaded the BMA to lower the level of the bags. A fewweeks later, communitymembers inspected the water gate at the border of the two districts where the water level was two meters higher on the Lum Luk Ka side. Their finding clearly suggests that BMA was protecting Sai Mai at Lum Luk Ka's expense. The community members successfully negotiated with BMA to open thewater gate by 50 cm. Both of thesemeasures led to a lower level of floodwater in their community (Kukot Pattana community leader, personal communication, September 14 2014).

5. Discussion

This paper has applied the politics of scale as a framework to examine how decentralization reforms and the associated power relations between government agencies at different levels affected disaster risk outcomes in Thailand, particularly during the 2011 floods in Central Thailand. It argues that decentralization has been incomplete due to the retention of power and resources by central bureaucrats, and the continued weak capacity of LAOs. Moreover, the country's overall fragmented and politically polarized gover- nance has hindered policy coherence at all levels, including the local level. Incomplete decentralization alongside persistent frag- mentation along ministerial and sectoral lines has undermined disaster governance, and distributed risks unevenly and unfairly. The governance of land and water proceeds in parallel, with inad- equate cross-scale links or multi-level collaboration. Furthermore, both the central government and LAOs have had to rely too much on flood protection infrastructure, as they have a long history of prioritizing rapid urban and industrial development and invest- ment in previously green spaces which have a high risk of floods (Marks, 2015), ensuring that the growth of many urban areas remain incompatible with DRR.

These governance weaknesses visibly materialized during the 2011 floods. Except for BMA, LAOs once again had insufficient ca- pacity to effectively respond to the floods, and were given insuffi- cient assistance by the central government. The central government sought to monopolize power, did not consult local communities, had limited capacity to enforce all its decisions, distributed risks unevenly, and overall, performed poorly. Similarly, the BMA dominated other much smaller local government units within and beyond its formal boundaries. At the same time, local communities contested the scalar configuration of the floods, and pressured the central government and BMA to more evenly distribute the floodwater.

Our findings suggest there are merits and limitations of decentralization in practice for DRR (Table 1). The case study

Table 1 The potential advantages and disadvantages of decentralization of DRR responsibilities t

Flood period

Advantages Disadvantages

Pre-event Understand local vulnerabilities Higher local engagement in preparedness

Inconsistent actions Unintended side effects of on other locations Inefficiency

Imminent Relevant warnings and action information reach local communities more

Inconsistent information

During Relief reaches more affected, and sooner Logistics incomplete Local resources insufficien

Post- event

Match local needs and capacities or resources Inefficient Incoherent Insufficient resources Weak coordination of ext inflow

suggests, however, that the potential advantages of decentraliza- tion for DRR have not been fully realized in Thailand, primarily because the central government has not sufficiently devolved po- wer and empowered LAOs. Even if LAOs better understood local vulnerabilities, without adequate human and financial resources it is difficult for them to conduct vulnerability reduction activities before disasters. The same can be said of the potential benefit of warnings at the local level. During the floods, LAOs not only failed to apply the national-level warning to the context of the area for which they were responsible, but also lacked sufficient funding and autonomy to provide relief and to reduce the level and length of the flooding. At the same time, some of the disadvantages of decen- tralization were evident, particularly insufficient resources and inconsistency, which weakened the governance of the floods.

The findings above suggest the need for more polycentric institutional architecture in Thailand which would overcome existing fragmentation problems. Combined with concurrent democratization, such a structure would likely distribute disaster risks and burdens more fairly and be able to respond to changing risks more effectively at multiple levels and scales. It would enable many of the advantages listed in the table to materialize, while diminishing the potential of these disadvantages from arising as well.

However, a few bottlenecks interwoven within the country's scalar politics currently impede this institutional architecture from developing. First, local elites have captured many LAOs, such as seen in Buriram, and consequently there is a lack of political accountability and transparency at the local level. Since it is difficult for citizens to observe the implementation and effects of preven- tative activitiesdsuch as improving building codes, early warning systems, and land use controldand since they instead tend to reward politicians for relief efforts that they can evaluate more easily, local politicians have favored relief and unregulated devel- opment over prevention. Second, powerful national-level agencies which, as expected, have been reluctant to concede power and collaborate with LAOs, have not been pushed enough to do so. In particular, central-level bureaucrats, including provincial governors and powerful line ministries, continue to wield significant power at the sub-national level. While there have been proposals frommany sides to devolve power, through measures such as electing gover- nors, they have not come to fruition. The recent coup and subse- quent re-centralization suggest that their powers are unlikely to wane soon. Third, the highly politicized wider political economy makes DRR coordination between different scales and levels more difficult. For instance, the in-fighting between FROC and BMA during the floods weakened the overall response. Fourth, the

o local authorities.

Interventions

preparations Incentives for improved coordination of land- and water-related planning Build capacities of LAOs in DRR

Improved sharing of early warning information and high-risk areas with local communities

t Improved coordination of disaster management agencies and LAOs

ernal aid

Incentives and better mechanisms to integrate DRR into normal LAO development plans and actions

D. Marks, L. Lebel / Habitat International 52 (2016) 57e66 65

country's administrative system is overly fragmented: too many agencies are responsible for similar tasks and responsibilities in one sector. The sheer number of agencies, particularly in the water sector, and their reluctance to work along different scales and levels, weakens collective action which is needed for polycentric DRR, especially during disaster events. Last, the political discourse of scale in communities and LAOs on one side and central author- ities on the other reveals that mistrust runs deep between the two. For example, while central government leaders claimed that local governments exacerbated the flooding by defying their orders, LAOs and many local communities blamed the central government or BMA for causing the floods.

Based on our findings, in Table 1 we suggest possible in- terventions to help overcome these bottlenecks and improve DRR in Asia during all of the phases of the disaster cycle. These sug- gested interventions will likely be difficult in many countries because they could challenge pre-existing scalar and interagency relations. However, research shows that structures of power gov- erning disasters are not static. Rather, because disasters can open political space and act as catalysts for change (Pelling & Dill, 2010), after disasters occur, the greatest opportunity arises for these suggested interventions to be successful and for scalar relations to change.

6. Conclusion

Little research has been conducted on how scalar politics affects disaster governance in Asia; this paper has addressed this gap. In particular, it has identified how Thailand's incomplete decentral- ization and administrative fragmentation has created numerous barriers to polycentric disaster governance. It has suggested some potential ways to improve decentralized disaster governance, in particular, through supporting local government capacities so they serve their local communities better. The politics of scale lens has also revealed how arguments frequently refer to scale and level, for example, when circumscribing boundaries around what should be protected, and where blame or responsibility lie in dealing with floodwater. More fundamentally, scalar arguments are used to both promote and resist efforts to decentralize power and administrative functions related to development and DRR. A comparison of these findings from a major flood event in Thailand with research con- ducted onmajor disasters in other Asian countries would be helpful to see whether similar bottlenecks arise, similar tools are used to contest uneven risk distribution, and differing biophysical condi- tions and forms of decentralization affected scalar relations. Such comparative research would also help test the proposition emerging from our analysis that more polycentric governance ar- rangements would benefit disaster risk reduction.

Acknowledgments

The authors would like to express thanks to John Connell and Richard Friend for their helpful comments and feedback and to the Stockholm Environment Institute's Asia Centre for hosting one of the authors during the time he conducted the interviews in Bangkok. Any remaining errors and omissions are ours alone.

References

Asia Development Bank. (2012). Thailand sector assessment (Summary): Water resource management (draft).

Bangkok Post. (2011, November 6). BMA gives Froc crisis ultimatum. http:// pattayatoday.net/news/thailand-news/bma-gives-froc-crisis-ultimatum/.

Barber, R. (2012). Responding to emergencies in Southeast Asia: Can we do better? Save the children Asia Regional office. Singapore.

Blackburn, S. (2014). The politics of scale and disaster risk governance: barriers to

decentralisation in Portland, Jamaica. Geoforum, 52(1), 101e112. http://doi.org/ 10.1016/j.geoforum.2013.12.013.

Brown, C. J., & Purcell, M. (2005). There's nothing inherent about scale: political ecology, the local trap, and the politics of development in the Brazilian Amazon. Geoforum, 36(5), 607e624. http://doi.org/10.1016/j.geoforum.2004.09.001.

Bulkeley, H. (2005). Reconfiguring environmental governance: towards a politics of scales and networks. Political Geography, 24(8), 875e902. http://doi.org/10. 1016/j.polgeo.2005.07.002.

CartoGIS Coordinator. “Bangkok Metropolitan area.” Canberra, Australia: College of Asia and the Pacific, The Australian National University. Accessed 01.09.15. http://asiapacific.anu.edu.au/mapsonline/base-maps/bangkok-metropolitan- area.

Chang Seng, D. S. (2013). Tsunami resilience: multi-level institutional arrange- ments, architectures and system of governance for disaster risk preparedness in Indonesia. Environmental Science & Policy, 29(1), 57e70. http://doi.org/10.1016/j. envsci.2012.12.009.

Chantornvong, S. (2000). Local godfathers in Thai politics. In Money and power in provincial Thailand. Honolulu: University of Hawaii Press.

Chau, V. N., Holland, J., & Cassells, S. (2014). Institutional structures underpinning flood management in Vietnam. International Journal of Disaster Risk Reduction, 10(Part A), 341e348. http://doi.org/10.1016/j.ijdrr.2014.10.008.

Department of Local Administration. (2010, January). Local government in Thailand. Presented at the delegation from the kingdom of Bhutan study visit to department of local administration, Bangkok, Thailand.

Djalante, R., Thomalla, F., Sinapoy, M. S., & Carnegie, M. (2012). Building resilience to natural hazards in Indonesia: progress and challenges in implementing the Hyogo framework for action. Natural Hazards, 62(3), 779e803. http://doi.org/10. 1007/s11069-012-0106-8.

Dufhues, T., Theesfeld, I., & Buchenrieder, G. (2014). The political economy of decentralization in Thailand: how past and present decentralization affects rural actors' participation. European Journal of Development Research. http:// dx.doi.org/10.1057/ejdr.2014.68.

Elinoff, E. (2013). Architectures of citizenship: Democracy, development, and the pol- itics of participation in Northeastern Thailand's Railway communities. PhD. San Diego: University of California, San Diego.

Faguet, J.-P. (2014). Decentralization and governance. World Development, 53(1), 2e13. http://doi.org/10.1016/j.worlddev.2013.01.002.

Frederickson, T. (October 27, 2011). Northern Run-off hits Sai Mai. Bangkok Post. http://www.bangkokpost.com/learning/easy/263441/northern-run-off-hits- sai-mai.

Friend, R., Jarvie, J., Reed, S. O., Sutarto, R., Thinphanga, P., & Toan, V. C. (2014). Mainstreaming urban climate resilience into policy and planning; reflections from Asia. Urban Climate, 7, 6e19. http://doi.org/10.1016/j.uclim.2013.08.001.

G€org, C. (2007). Landscape governance: the “politics of scale” and the “natural” conditions of places. Geoforum, 38(5), 954e966. http://doi.org/10.1016/j. geoforum.2007.01.004.

Guerrin, J., Bouleau, G., & Grelot, F. (2014). “Functional fit” versus “politics of scale” in the governance of floodplain retention capacity. Journal of Hydrology, 519(Part C), 2405e2414. http://doi.org/10.1016/j.jhydrol.2014.08.024.

Harries, T., & Penning-Rowsell, E. (2011). Victim pressure, institutional inertia and climate change adaptation: the case of flood risk. Global Environmental Change, 21(1), 188e197. http://doi.org/10.1016/j.gloenvcha.2010.09.002.

Houdret, A., Dombrowsky, I., & Horlemann, L. (2014). The institutionalization of River Basin management as politics of scale e insights fromMongolia. Journal of Hydrology, 519(Part C), 2392e2404. http://doi.org/10.1016/j.jhydrol.2013.11.037.

Johannessen, Å., & Hahn, T. (2013). Social learning towards a more adaptive para- digm? Reducing flood risk in Kristianstad municipality, Sweden. Global Envi- ronmental Change, 23(1), 372e381. http://doi.org/10.1016/j.gloenvcha.2012.07. 009.

Kokpol, O. (2011). Decentralization process in 1990e2010: in case of Thailand. In Presented at the 1st meeting of research committee of decentralization in Asian countries, Tokyo. Retrieved from http://kpi.ac.th/media/pdf/M10_141.pdf.

Kritsanaphan, A., & Sajor, E. (2011). Intermediaries and informal interactions in decentralised environmental management in peri-urban Bangkok. International Development Planning Review, 33(3), 247e272. http://doi.org/10.3828/idpr.2011. 11.

Lebel, L., Garden, P., & Imamura, M. (2005). The politics of scale, position, and place in the governance of water resources in the Mekong region. Ecology and Society, 10(2), 18.

Lebel, L., & Sinh, B. T. (2009). Risk reduction or redistribution? Flood management in the Mekong region. Asian Journal of Environment and Disaster Management, 01(01), 25. http://doi.org/10.3850/S179392402009000040.

Lebel, L., Sinh, B. T., Garden, P., Seng, S., Tuan, L. A., & Truc, D. V. (2009). The promise of flood protection: Dikes and dams, drains and diversions. In Contested waterscapes in the Mekong region: Hydropower, livelihoods, and governance (pp. 282e306). Earthscan.

Litvack, J., Seddon, J., & Ahmad, J. (1999). Decentralization briefing notes. Washington, DC: World Bank Institute.

Maneepong, C., & Webster, D. (2008). Governance responses to emerging peri- urbanisation issues at the globalelocal nexus: the case of Ayutthaya, Thailand. International Development Planning Review, 30(2), 133e154. http://doi. org/10.3828/idpr.30.2.3.

Manuta, J., Khrutmuang, S., Huaisai, D., & Lebel, L. (2006). Institutionalized in- capacities and practice in flood disaster management in Thailand. Science and Culture, 72(1/2), 10.

D. Marks, L. Lebel / Habitat International 52 (2016) 57e6666

Marks, D. (2015). The urban political ecology of the 2011 floods in Bangkok: the creation of uneven vulnerabilities. Pacific Affairs, 88(3).

Molle, F. (2007). Scales and power in river basin management: the Chao Phraya River in Thailand. The Geographical Journal, 173(4), 358e373.

Nagai, F., Funatsu, T., & Kagoya, K. (2008). Central-local government relationship in Thailand. In Local government in ThailanddAnalysis of the local administrative organization survey. Chiba, Japan: Institute of Developing Economies.

Nagai, F., & Kagoya, K. (2008). Local capability and decentralization in Thailand. In Local government in ThailanddAnalysis of the local administrative organization survey (pp. 51e60). Chiba, Japan: Institute of Developing Economies.

Pahl-Wostl, C., Lebel, L., Knieper, C., & Nikitina, E. (2012). From applying Panaceas to mastering complexity: toward adaptive water governance in river basins. Environmental Science & Policy, 23, 24e34. http://dx.doi.org/10.1016/ j.envsci.2012.07.014 (November).

Pelling, M., & Dill, K. (2010). Disaster politics: tipping points for change in the adaptation of sociopolitical regimes. Progress in Human Geography, 34(1), 21e37. http://doi.org/10.1177/0309132509105004.

Reynolds, C., &, Team. (2012). Time's arrow and the Burden of the past: a primer on the Thai Un-State. Sensate, 1e16.

Sato, J. (2011). State inaction in resource Governance: Natural resource control and bureaucratic oversight in Thailand (No. 36). Tokyo: JICA Research Institute.

Satterthwaite, D. (2013). The political underpinnings of cities' accumulated resil- ience to climate change. Environment and Urbanization, 25(2), 381e391. http:// doi.org/10.1177/0956247813500902.

Shatkin, G. (2004). Globalization and local leadership: growth, power and politics in

Thailand's Eastern Seaboard. International Journal of Urban and Regional Research, 28(1), 11e26. http://doi.org/10.1111/j.0309-1317.2004.00500.x.

Skidmore, M., & Toya, H. (2013). Natural disaster impacts and fiscal decentralization. Land Economics, 89(1), 101e117.

Tangwisutijit, N. (2007). Bangkok will be underwater, experts predict. The Nation. Thomalla, F., & Larsen, R. K. (2010). Resilience in the context of tsunami early

warning systems and community disaster preparedness in the Indian Ocean Region. Environmental Hazards, 9(3), 249e265.

UNISDR. (2005). Hyogo framework for action 2005e2015: building the resilience of nations and communities to disasters. In Extract from the final report of the World Conference on Disaster Reduction (A/CONF. 206/6).

Wheater, H., & Evans, E. (2009). Land use, water management and future flood risk. Land Use Policy, 26(Suppl. 1), S251eS264. http://doi.org/10.1016/j.landusepol. 2009.08.019.

Wijitbusaba, A. M. (2013). Urban risk and vulnerabilities of coastal Megacity of Bangkok, Thailand. In Proceedings of the Resilient Cities 2013 congress. Bonn, Germany.

Wisner, B., Blaikie, P., Cannon, T., & Davis, I. (2004). At risk: Natural hazards, people's vulnerability and disasters (2nd ed.). London; New York: Routledge.

Wongpreedee, A. (2007). Decentralization and its effect on provincial political po- wer in Thailand (The front of area studies). アジア・アフリカ地域研究 ¼ Asian and African Area Studies, 6(2), 454e470.

World Bank. (2012). Thailand e Public financial management report 2012: Improving service delivery: Overview (No. 67486) (pp. 1e58). Washington D.C.: The World Bank.