Development Policy Review, 2016, 34 (2): 253--276
Dangers of Decentralisation in Urban Slums: A Comparative Study of Water Supply and Drainage Service Delivery in Kolkata, India Indranil De and Tirthankar Nag* Clientelism may lead to the underprovision of services which are deemed suitable for decentralisation. Water distribution and drainage services, managed from a lower level of municipal authority, are liable to be affected by clientelism and consequent underprovision. Water quality, maintained from a higher municipal layer, is not likely to be affected by clientelism. Capture by politically influential and dominant social and religious groups is likely to take place for important services like water supply. The article suggests that awareness, measurability, importance and resource intensiveness of service are additional factors to be considered for assessing the suitability of a sector for decentralisation. Key words: Decentralisation, political economy, urban, slum, India
1
Introduction
Inadequacy of and dissatisfaction with service delivery are quintessential problems for the urban poor in India. Decentralisation, introduced by the 74th Constitutional Amendment, aimed at increasing the efficiency of service provision and people’s participation in urban governance. However, it has failed to improve the basic amenities in urban slums. Provision of basic services, including water and sanitation, by local governments in developing countries is hampered by the low capacity of local governments, corruption, elite capture and political influence (Bardhan 2002; Bardhan and Mookherjee, 2000; Asthana, 2003; Slaymaker and Newborne, 2004; Mtisi and Nicol, 2003). Scholars have also expressed concern over the increasing inequity that has accompanied decentralisation (Prud’homme, 1994; Litvack et al., 1998). Prud’homme (1994: 27) has spelt out a few criteria ‘to find out which services or sectors would lend themselves more easily to decentralization’. This article argues that these criteria are inadequate and hence the dangers of decentralisation are more profound. It compares the delivery of water supply and drainage services in slums of Kolkata, which is the state capital of Indian state of West Bengal. It finds that the greater the suitability of service for decentralisation, the higher the chance of underprovision. Moreover, the less suitable the service for decentralisation, the greater the chance of capture by politically influential and socially dominant groups.
*Assistant Professor, Institute of Rural Management Anand, Anand, Gujarat, India 388001 (
[email protected]); Associate Professor, International Management Institute, Kolkata, West Bengal. © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
254
Indranil De and Tirthankar Nag
The urban water supply in India faces many problems. These include huge gaps in water demand and supply, poor operation and maintenance of water supply systems, huge water losses mainly caused by leakages in transmission and distribution lines, a significant proportion of non-revenue and unaccounted flow of water, striking intra-urban disparities, inadequate supply of safe drinking water to poor communities, insufficient supply quantities, intermittent supply, poor water quality, low tariff for domestic connection/inappropriate pricing, lack of attention given to rationalisation of tariffs, and low cost recovery, among others (IDFC, 2011, Aijaz, 2010; McKenzie and Ray, 2009). Misra and Goldar (2008) report that total dissolved solids and alkalinity are higher than permissible limits in a sample collected from one of the treatment plants in Delhi. Political manipulation and deliberate overlooking of water supply add to the problem as observed in Mumbai, where services are underprovided to the poor and the socio-politically neglected Muslim population (Contractor, 2012). The present article focuses mostly on the political economy of water and drainage service delivery and makes an attempt to identify an additional set of criteria to judge the suitability of a sector for decentralisation in developing countries. Prud’homme (1994) posits that decentralisation can be dangerous by exacerbating disparity, allocative and production inefficiency and corruption; but he did not address issues relating to the impact of political competition or the possibility of clientelism, and consequent underprovision of services. Political or electoral competition acts as one of the main mechanisms through which democracy engenders accountability and responsiveness. Competitive elections incentivise politicians to respond to the needs of their constituents (Skilling and Zeckhauser, 2002; Lake and Baum, 2001; Holbrook and Van Dunk, 1993). Nevertheless, empirical evidence concerning the relationship between electoral competition and the provision of basic services is not uniform. An attempt has been made to explain the negative relationship between political competition and government response through the question of ‘clientelism’, which leads to an excessive tendency on the part of political patrons to provide private goods to their ‘clients’. Clientelism is a system of exchange based on political subordination in exchange for the discretionary granting of available public resources and services (Heredia, 1997; Kitschelt and Wilkinson, 2007). Lemarchand (1972: 69) explains clientelism as a ‘more or less personalized relationship between actors or sets of actors commanding unequal wealth, status or influence, based on conditional loyalties and involving exchange transactions’. It is a market idea which perceives neighbourhood organisations as consumers and the municipality as a producer of services (GarcıaGuadilla and Perez, 2002). Clientelism derives from a decentralisation process oriented towards deconcentration and privatisation which provides limited credibility to political competitors in the face of high political competition and political fragmentation (Garcıa-Guadilla and Perez, 2002; Keefer, 2002; Gay, 1990). Garcıa-Guadilla and Perez (2002) opine that in the Chacao municipality in Venezuela clientelism developed due to distortions in the process of transition from representative democracy to participatory democracy. A study of El Alto, Bolivia by Lazer (2004) suggests that clients barter their support during electoral campaigns, in exchange for © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
255
individual and collective gain, as a strategy to redress the ignoring of the poor by bureaucracy and to make politicians more representative and responsive. Based on his study of the Vila Brasil slum near Rio de Janerio, Gay (1990) elucidates that clientelism is a logical and rational choice for receiving collective benefits given existing social, economic and political inequality and an exclusionary national political system. Benit-Gbaffou (2011) found that in Johannesburg clientelism allowed for the local and immediate accountability of politicians, but hindered the contestation of existing policies and dominant power structures, thereby undermining democracy Clientelism tends to reduce the provision of basic services. With an increase in the number of clients or individuals it is difficult for political agents to make promises about public service provision to large segments of voters (Keefer and Khemani, 2004). In socially and ethnically fragmented societies, governments tend to focus spending on areas of interest to specific individuals or clients, which, in turn, leads to less spending on public goods. Chanie (2007) found that the clientelist relationship between central and regional leaders led to the failure of decentralisation to achieve allocative and production efficiencies, accountability and responsiveness in Ethiopia. Delivery of basic services improves with people’s participation. In Uruguay, Montevideo’s public services improved through institutional reforms by making governance less centralised, bureaucratic and clientelist (Canel, 2001). Local participation in municipal affairs, and the creation of democratic citizenship at community level, developed skills and leadership. Along similar lines, Chattopadhyay (2012a) found that in West Bengal municipal services, including water supply, drainage, garbage collection and road maintenance, are generally poor. However, these services are better in municipalities where an organic link with the citizens through the involvement of local people in decision-making and monitoring has been established. In general, it is difficult to establish due to co-option in ward committees and irregularities in monthly meetings and annual general meetings of ward committees (Chattopadhyay, 2012b). Prud’homme (1994) puts forward several criteria that determine the suitability of the sector with regard to decentralisation. These include externality (negatively related), chargeability (positively related) and technicity (negatively related). Following these criteria, drainage is more suitable for decentralisation than water supply. However, studies have failed to identify that the services deemed important and resource intensive are more likely to get captured by influential and dominant groups. Services are susceptible to clientelism depending on the measurability of service standards and awareness of the service. The study consists of seven sections. Section two puts forward the theoretical argument for clientelist behaviour on the part of political agents in the context of delivery of public services. The sampling plan of the study has been illustrated in section three. The next section discusses the economic and political profile of selected households. The methodology of analysis, including the description of study variables, is discussed in section five. The regression results have been illustrated in section six, while section seven summarises the results and makes concluding comments. © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
256
2
Indranil De and Tirthankar Nag
Political clientelism and service delivery
The present study is centred on slums under the jurisdiction of the Kolkata Municipal Corporation (KMC). The Corporation of Calcutta was established under a Royal Charter in 1726. The present structure of KMC was formulated on the basis of the Calcutta Municipal Corporation Act of 1980. It lays down a Mayor-in-Council form of government comprising a Mayor, Deputy Mayor and a maximum of ten elected members. The Mayor-in-Council interacts with the Borough Committees, which consist of several wards. The Borough Committees place claims for public services to the Mayor-in-Council, and this furnishes the commensurate funds and policies. The Mayor-in-Council, in turn, places claims to the state government which provides funds and policy to the Mayor-in-Council (Chandra, 2004). The Borough Committees consist of a legislative structure and an executive structure. The legislative structure comprises the Borough Chair, who is elected from the Ward councillors. The executive structure contains executive engineers, health officers and other government officials. It also has lower-level government staff. Borough Committees are directly responsible for the maintenance of local services like water supply and drainage. The Ward Committees are established in each ward under the West Bengal Municipal Corporation Laws, 1994. Ward Committees have the authority to raise ward-level problems, assign priorities and take responsibilities for the supervision and oversight of various activities with active popular participation. In a major step towards decentralisation the Calcutta Municipal Corporation (Amendment) Act of 2001 set the ground for the establishment of Area Sabhas in every ward comprising one or more polling booths. These consist of individuals whose names are included on the electoral roll of the polling booth concerned. There are non-hierarchical relationships between all these institutions, although relationship strengths are not equal. Relationship strength is high, however, between the Ward Committee and the Borough Committees, the Borough Committees and Mayor-in-Council, and the Mayor-in-Council and state government. In Kolkata slums, the grievances of slum dwellers regarding basic services are registered at the Bustee (‘slum’ in the local language) level as a part of Bustee Services. Government officials and elected representatives have low (perceived) credibility vis- a-vis Bustee services. Das (2009) found that Ward Councillors or KMC officials expressed indifference or lack of interest regarding problems surrounding water supply. They also demonstrated lack of promptness when it came to acting on problems of sanitation, drainage and public hygiene. Therefore, by channelling the deficit through provision of better services at Bustee level to patronbased groups, clientelism can manipulate the distribution of water and drainage services. However, quality of water is likely to remain unaffected by a patronagebased party–voter relationship. This is because water is treated at the treatment plant and water supply is provided through a network of channels. These activities are monitored by the Borough Committees. Inefficiencies occurring at higher levels of local government are likely to be more easily identified as compared to inefficiencies occurring at the Bustee level. Information regarding inefficiencies may have negative consequences for the Borough Committees and its chairman. Studies © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
270
Indranil De and Tirthankar Nag
diluted once we consider the other communities in a jurisdiction. The chances of good quality water fall by up to 60% where the local councillor is Muslim. Furthermore, the probability of getting good quality water is lower by at least 18% in Central Kolkata slums that are Muslim dominated. Local government appears to play a major role in maintaining the quality of water supply. Quality of water improves if the current councillor happens to have won elections with a higher percentage of votes and a higher and improved margin as compared to the preceding elections; that is to say, if political competition reduces. The higher the political fragmentation (measured by the number of contestants per ten thousand voters), the better the chances of good quality water. Wards characterised by both high competition and high fragmentation have higher chances of receiving good quality of water supply. The probability of getting good quality water increases by 32% if the councillor is affiliated to the AITC. The chances of getting good quality water increases by up to 23% if the local government is made aware of the feedback of households on quality of water. Hence, organic linkage with local government improves services. Furthermore, if the local councillor is a repeat (re-elected) councillor of the previous municipal council, then the chances of getting good quality water fall by 61%. This could be due to less effort on the part of the councillors in governance. Overall, the results illustrate the fact that political competition has a negative impact, political fragmentation has a positive impact, and the combination of political competition and fragmentation has a positive impact on the quality of water.
6.3 Drainage The maximum likelihood estimates of the determinants of ‘good_drainage’ are illustrated in Table 10. The coefficients of NK, CK, SCST, MC, and KEIP are positive while the coefficients of PL, CONTEST, AITCC, VOTE, MARGIN, CMARGIN and HFLM are negative. All independent variables are statistically significant. SCST and MC are significant in all the models and hence robust. With respect to the overall model fit, all three Tobit models appear to be statistically significant. Slums located on private land are less likely to get services like cleaning of drainage within a fortnight. Households from North and Central Kolkata slums stand good chances of drainage cleaning within a fortnight. This could be due to high population density in these regions. Poorer caste households have higher chances of receiving good drainage services. Moreover, households represented by a Muslim councillor have higher chances of good drainage services. Candidates victorious in more highly competitive elections (lower percentage of votes, less margin and reduced margin as compared to previous elections) demonstrate better chances of good drainage service delivery. However, political fragmentation (measured by number of contestants per ten thousand voters) reduces the chances of better service delivery. Moreover, high political competition coupled with high political fragmentation is likely to lead to lower delivery of services. Councillors not associated with the majority party might try to showcase their work by delivering services such as drainage, which is under their direct control. The Kolkata © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
258
Indranil De and Tirthankar Nag
supply is delivered from a higher level of local government covering a large section of the population through network channels. Treatment of water is carried out in plants responsible for the distribution of water to several wards. Hence, quality of water has implications for households spread across several wards. The reservation support for quality is expected to vary across households as awareness about quality of water and possible alternatives is likely to vary across individuals and communities. However, physical access and distribution (volume) of water is decided at the local level. Drainage services are also managed from lower levels of local government catering to a small section of the population. The aggregate demand for quantity of water and drainage services takes into account fewer individuals within a single ward. These individuals face similar situations in terms of provision and alternatives. Hence, variation of reservation support is likely to be narrower for quantity of water and drainage service vis- a-vis the quality of water supply. Therefore, the aggregate demand for water quality would be much more elastic compared to the aggregate demand for quantity of water and public drainage services. High electoral competition impels the incumbent to reduce minimum political support for each unit of public service in an effort to lure away supporters. In other words, given higher electoral competition, an incumbent will be willing to provide higher and/or better public service at the same level of support. However, the incumbent can provide higher/better public service only if it is feasible. Higher political competition curbs the bargaining power and clout of the incumbent at higher levels of local government. It reduces their share of financial resources, technical and non-technical manpower and public goods, like water. Financial resources and technical manpower are crucial for maintaining both the sufficiency and the quality of the water supply. We have considered two cases in this regard. Case I deals with the market situation where incumbents do not face any constraint in the delivery of higher/better service. Case II explains the market adjustments where incumbents face constraints in the delivery of service. Case I: Feasibility constraint not binding. The final objective of the incumbent is to increase total support. Due to the fall in minimum political support required for public service, the equilibrium level of support will fall and the level of public service will increase. In the context of water supply quality, where the aggregate demand for the service is elastic, total support increases as equilibrium political support for the incumbent falls. As total political support rises, it becomes an acceptable equilibrium for the incumbent. Whereas, owing to the inelastic aggregate demand for quantity of water supply and drainage, a fall in the equilibrium support is expected to lead to a fall in total support, which is detrimental to the incumbent. To avoid this fall the incumbent raises political support for provisioning public service, thereby garnering higher total support. The incumbent provides service to fewer individuals willing to offer higher political support. At higher required political support, the overall level of public service decreases. This higher political support leads to the underprovisioning of public service. This disequilibrium is economically inefficient, being driven by political clientelism. While the incumbent is better off in terms of total political support with this disequilibrium, this is only possible if individuals willing to provide support form © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
259
one or more groups. This is because public service can only be provided to supportive individuals if they reside within a compact geographical boundary. In other words, individuals need to form collective support groups, with little dissent to the offer of support. The higher the political fragmentation, the smaller the size of the group and the greater the chances of the formation of such small supportive groups. Theories of collective action suggest that social norms developed through face-to-face communication can produce substantial and sustainable co-operation (Ostrom, 2000). Any norm to support a political party may be developed in a slum where individuals live in close vicinity. Moreover, such groups should have a ‘critical mass’ for producing collective action (Oliver et al., 1985). Public service of the nature discussed in this study also needs to have a critical size in order to be delivered. The level of group support ultimately decides the amount of public service to be delivered in the disequilibrium. Political parties are likely to foster political support until it becomes prohibitive and reaches the elastic portion of the demand for public service curve. The demand and supply function of public service is illustrated below @s \0 @q
. . .. . .. . .Inverse demand function of public service
@s [0 @q
. . .. . . . . .Inverse supply function of public service
s ¼ D (q); s ¼ S (q);
where s is the political support and q is public service. The initial equilibrium public support and quantity of public service would be s0 and q0. In the face of high political competition, the political support required per unit of public goods provision falls. Therefore, the supply curve of public service shifts to the right from S0 to S1. The new equilibrium is at s1 and q1. At this new equilibrium, political support would come down (s1 < s0) and provision of public good would increase (q1 > q0) as illustrated in Figure 1. In the presence of clientelism, the public service and required support would be settled at (s2 and q2). Higher competition reduces the availability of public service from Q0 to Q1. The feasibility constraint is not binding on equilibrium (s1 and q1). Case II: Feasibility constraint is binding. The limited availability of public service may place constraints on public service delivery, both initially and later when political competition increases. If the feasibility constraint is binding on the initial equilibrium (s0 and q0) then the public service may be undersupplied. Figure 2 illustrates that if the initial constraint is at Q00 then the initial equilibrium (s3, Q00 ) is inefficient. However, such inefficiency does not arise where the constraint is at Q0. An increase in political competition may further tighten the constraint to Q1. The equilibrium shifts to (s4, Q1) if Q1 is binding. This equilibrium is also economically inefficient. In this case, higher political competition leads to lower delivery of public service. The incumbent is likely to distribute the service to individuals providing the
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
260
Indranil De and Tirthankar Nag
highest support. Political clientelism of this kind is possible when incumbents determine the distribution of limited public service.
Figure 1: Market for public service (feasibility constraint not binding)
3
Sampling plan
The slums and households have been selected through a mixture of cluster sampling and systematic sampling techniques. After initially segregating Kolkata slums into Central, North and East, and West and South regions, they have been further segregated by overlaying the population and average monthly income of slums.1 The final selection of slums from these stratums has been done randomly. A total of 23 slums, each from different wards, has been selected, both from notified and nonnotified areas (Table 1).2 The households have been selected from the slums by systematic sampling method. A total of 541 households was selected; the number of
1. Kolkata Municipal Corporation (KMC) provided information on population and monthly average income of slum dwellers. 2. Notified slums in India are notified as slums by state or local government under any Act. The KMC Act, 1980 defines these slums as ‘an area of land not less than 700 square metres occupied by, or for the purposes of, any collection of huts or other structures used or intended to be used for human habitation’. These are government authorised slums and enjoy security of tenure to a large extent. Slums in the non-notified areas have been developed mainly by encroaching on government or railway land. These slums do not have any security of tenure as they are unauthorised. © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
261
selected households in a slum is proportionate to the total population of the selected slums. The distribution of the slum population per region is displayed in Table 2.
Figure 2: Market for public service (feasibility constraint binding)
Table 1: Number of slums selected in the survey Central
North and East
West and South
All
6 1 0 1 7
7 2 2 0 9
6 1 1 0 7
19 4 3 1 23
Notified/ Private Land Non-notified Government Land Railways Land Total
Table 2: Distribution of sample households by region
Central North and East West and South Total
No. of Households
No. of Households (%)
Non-notified (%)
157 104 280 541
29 19 52 100
18 18 1 9
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
262
4
Indranil De and Tirthankar Nag
Demographic, economic and political proligfiligle of the selected households
The population density in the slum areas of selected wards is highest in Central Kolkata and lowest in West and South Kolkata. The overall population density is 66,193 per square km. This is much higher than the population density of 24, 252 per square km of the Kolkata district. Overall, Muslims make up the majority in Kolkata slums, accounting for more than 50% of slum households. However, the concentration of social and religious groups varies across regions. Muslims constitute a significant majority in Central Kolkata. Hindu General (mainly upper castes) and Other Backward Class (a collective term used by the Constitution and Government of India to classify castes which are socially and educationally disadvantaged, here including the historically most disadvantaged Scheduled Castes (SC) and Scheduled Tribes (ST) households) are more prevalent in North and East Kolkata slums (Table 3). The distribution of households by economic condition, expressed by monthly per capita consumption expenditure (MPCE), indicates that the MPCE of 8% of sample households fall below Rs. 830.6, which the Planning Commission of India gave as the poverty line for urban West Bengal for 2009–10. As Table 4 shows, the percentage of households below the poverty line for Kolkata as a whole is highest in North and East Kolkata slums, and within Muslim households across religion and caste groups. However, the highest specific incidence of poverty is observed among the Hindu General category in North and East Kolkata (around 14%). The literacy rate of sample household members appears to be high at 74%. The percentage of literate slum dwellers is the highest, at 84%, in Central Kolkata slums, while it is lowest in the West and South Kolkata slums at 68%.
Table 3: Distribution of households by religion and caste (%)
Hindu General SC & ST Muslim Total
Central
North and East
West and South
All
25 3 72 100
48 17 35 100
35 11 54 100
34 10 56 100
Table 4: Households below poverty line (%)
Hindu General SC & ST Muslim All
Central
North and East
West and South
All
0 0 7 5
14 3 9 11
3 4 11 8
6 4 10 8
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
5
263
Methodology
Sufficiency of water supply has been measured by household response to the availability of water supply. Availability of water supply has been categorised as sufficiency throughout the year; sometimes sufficient or insufficient, and insufficient throughout the year. The taste, smell and colour of water have been considered as attributes of water quality. All these characteristics have been ordered by rank and the final quality index has been obtained by adding the ranks (Table 5). Our quality index ranges from rank 3 to 9. These ranks reveal the taste, smell and colour of water. Quality is good if all three characteristics of quality are good. Quality is deemed bad if at least one of these characteristics is bad. For the intermediary values of quality index, quality is considered as medium (Table 6). The quality of drainage service is categorised as good if the drains are cleaned on a fortnightly basis. If the drains are not cleaned within a fortnight then the quality of drainage service is deemed bad. Insufficient water supply has been reported by around 41% to 42% of sample households in all the regions. The intra-regional disparity is stark in West and South Kolkata where 19% of Hindu general caste households, as against 64% Muslim households, revealed insufficiency. In Central Kolkata, water supply is sufficient for more than 60% Hindu general caste and SC and ST households, as against a meagre 33% of Muslim households. The quality of water supply is good for 72% and bad for only 6% of the sample households. Households revealed bad water quality primarily in Central Kolkata (10% households). Bad water quality revealed by SC and ST households (15%) is highest amongst the religion and caste groups. Drainage in the slums of Kolkata is particularly bad, as around 9% of households
Table 5: Quality index = taste + smell + colour Taste Smell Colour
1, if bad 1, if bad 1, if bad
2, if medium 2, if medium 2, if medium
3, if good 3, if good 3, if good
Table 6: Water and drainage quality Water quality
Drainage quality
Good Bad Medium Good Bad
Taste and smell and colour good If taste and/or smell and/or colour bad Otherwise If cleaned once a fortnight Otherwise
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
264
Indranil De and Tirthankar Nag
have no access to drainage and 76% of households reported that drains were not cleaned in over a fortnight. In section 6 we have attempted to identify the determinants of sufficiency and quality of water supply as well as the quality of drainage service. The dependent variables include sufficiency of water supply throughout the year (sufficient), good quality of water supply (good_water) and good quality of drainage service (good_drainage). These qualitative dependent variables are binary by nature. Therefore, we have used limited dependent variable models to arrive at estimates of their determinants. We have constructed a Logit model for the determinants of sufficiency and quality of water supply. The quality of drainage service is a censored variable as many of the sample households do not have physical access to drainage and, thus, there is no issue of getting any drainage service. Hence, we have used the Tobit model to estimate the determinants of good quality drainage service. The independent variables in these models include type of land title, regional location of slum, ethnic identity of slum dweller as well of councillor, political affiliation of councillor and measures of political competition and fragmentation (Table 7). The higher the political fragmentation (higher value of ‘CONTEST’), the lower the service provision. It is expected that in wards identified as High Fragmentation and Low Margin (HFLM) are affected by political clientelism, and hence services are underprovided. The variables ‘VOTE’, ‘MARGIN’ and ‘CMARGIN’ represent political competition. The higher the political competition (lower value of ‘VOTE’, ‘MARGIN’ and ‘CMARGIN’), the higher the accountability of candidates winning municipal elections. However, higher political competition may lead to underprovision of public service if total public services are short in supply, as discussed in case II in section 2. In cases of total public service constraint, councillors are required to take the lead and exercise higher control in the Borough Committee (or higher levels of local government) for extra services. Therefore, the lower the competition the higher the political clout and bargaining power, and the better the chances of obtaining a larger share of public service. Moreover, a higher percentage of votes and higher margins can elevate the position and reinforce the political clout of the elected councillors. Political clout can be useful in services like water, which requires distribution of scarce resources across wards. AITCC is expected to have a positive impact on service delivery as AITC is the majority party in the corporation and therefore can have higher clout and control over higher levels of local government. On the other hand, Muslim councilors (MCs) are expected to be less able to exert pressure on local government to deliver more services due to the marginalisation of minorities and disadvantaged castes in local governance. This attitude extends to the household level, and is expected to lead to a negative relationship of MUSLIM and SC/ST regarding service delivery. Feedback (FEED) by slum dwellers regarding the quality of water supply to an agent of the local government is expected to lead to better quality water supply. Re-elected candidates (REELEC) are expected to have less incentive for better delivery of services. Other variables in the models are used as control variables. © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
265
Table 7: Independent variables Variable
Coding
Land title
NN = 1 if non-notified area, 0 otherwise PL = 1 if private land, 0 otherwise RL = 1 if railway land, 0 otherwise GL = 1 if government land, 0 otherwise CK = 1 if Central Kolkata, 0 otherwise NK = 1 if North Kolkata, 0 otherwise KEIP = 1 if ward covered by Kolkata Metropolitan Improvement Project,a 0 otherwise SCST = 1 if SC or ST household, 0 otherwise MUSLIM = 1 if Muslim household, 0 otherwise AITCC = 1 if councillor affiliated to All India Trinamool Congress, 0 otherwise VOTE = percentage of vote obtained by the winning candidate in last (2010) election MARGIN = percentage difference of vote between elected councillor and the runner-up (margin) in the last election CMARGIN = change in margin in 2010 over 2005 HFLM = 1 if wards have high competition / low margins (<20% difference with the runners-up) and high political fragmentation (>3 contestant per 10,000 votes polled), 0 otherwise CONTEST = number of contestants in the 2010 municipal elections per 10,000 electorate MIG = 1 if household migrated within last ten years, 0 otherwise MPCE = monthly per capita consumption expenditure (Rs) FEED = 1 if household believes that local government is aware of water quality, 0 otherwise REELEC = 1 if re-elected councillor, 0 otherwise
Location and project coverage of wards
Ethnic identity Political affiliation Political competition
Political fragmentation Migration status Economic condition Other variables
Notes: a) Kolkata Environmental Improvement Project (KEIP) is a multi-agency endeavour to arrest environmental degradation. Its work is mainly in the outer areas, including slums of the city, where the sewerage and drainage infrastructure is grossly inadequate and the drainage canals are choked by silt.
We have considered several models for each dependent variable in order to avoid multicollinearity from correlation between independent variables. The significant independent variables found across all the models are robust.
6
Regression results
6.1 Sufficiency of water supply The maximum likelihood estimates of the determinants of ‘sufficient’ have been presented in Table 8. We have constructed seven models. Results demonstrate that coefficients of GL, CONTEST, AITCC, MARGIN and CMARGIN have positive sign; and the coefficients of RL, MUSLIM, MEM, MIG, MPCE, MC, VOTE, CK © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
High Fragmentation & Low Margin (HFLM)
Change of Margin (CMARGIN)
Margin (MARGIN)
% vote obtained (VOTE)
AITC Councillor (AITCC)
Per capita consumption expenditure (MPCE) Election fought per 10,000 electorate (CONTEST) Muslim Councillor (MC)
Migrated (MIG)
No. of Members (MEM)
Muslim Household (MUSLIM)
Central Kolkata (CK)
Government Land (GL)
Railway Land (RL)
Dependent variable: sufficient quantity (sufficient_quantity)
0.29*** (26.7)
0.04*** (12.39) 0.12*** (6.84)
0.52#*** (21.19) 0.09* (3.01)
Model I
0.00008* (3.88)
0.22*** (23.53) 0.04*** (12.55)
0.26** (5.83) 0.33* (6.23)
Model II
0.99*** (9.54)
0.27*** (20.48)
0.16*** (10.8) 0.03*** (10.75)
0.39*** (9.03)
Model III
0.20** (5.02)
0.28*** (18.23) 0.27*** (29.38)
0.18*** (15.13) 0.03*** (8.29)
0.51*** (46.92)
Model IV
0.28*** (23.30)
0.03 (1.11)
0.23*** (21.24) 0.02** (8.36)
0.51*** (16.58)
Model V
Table 8: Determinants of Availability of Sufficient Water Supply
0.78*** (12.04)
0.32*** (34.34)
0.18*** (14.59) 0.03*** (11.16)
0.45*** (23.23)
Model VI
0.11* (3.13)
0.26*** (21.53)
0.17*** (12.96) 0.04*** (12.32)
0.34*** (8.12)
Model VII
266 Indranil De and Tirthankar Nag
312.9 67.4 64.54 0 0.09 513
313.135 67.4 64.08 0 0.0928 513
Log likelihood Percentage correct LR chi2 Prob > chi2 Pseudo R2 No. of observations
305.18 68 80 0 0.1159 513
Model III 297.85 69 94.65 0 0.14 513
Model IV 304.35 70.8 76.38 0 0.1138 513
Model V
303.84 67.4 82.67 0 0.12 513
Model VI
309.04 65.5 72.27 0 0.10 513
Model VII
Note: Numbers in parentheses are Wald chi-square statistics. #Marginal Effects, not the coefficients. *** indicates significance at 1% level, ** at 5% level and * at 10% level.
Model II
Model I
Dependent variable: sufficient quantity (sufficient_quantity)
Table 8: Continued
Dangers of decentralisation in urban slums
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
267
268
Indranil De and Tirthankar Nag
and HFLM have negative sign. All the independent variables barring CONTEST (in Model V) are statistically significant. GL and MEM are significant in all five models and hence robust. MUSLIM is also significant in four models. All seven models have been able to make more than 65% correct predictions. The availability of water depends on the title of land on which the slum is located. There is a 33% to 52% higher probability of getting sufficient water in Government land existing in non-notified areas. There is a 26% lower probability of acquiring sufficient water on the land encroached from the railways. Migration reduces the chances of getting sufficient water supply by 12%. The higher the percentage of the vote obtained by the councillor in municipal elections, the greater the chances of getting sufficient water. More convincingly, when the councillor wins the last elections – depicted by the higher margin and change in margin – the higher the chances of sufficient water supply. These imply that political competition reduces the chance of getting sufficient water. Political fragmentation (measured by number of contestants per ten thousand voters) has a positive but insignificant impact on sufficiency. Wards with both high competition and high fragmentation in the last election demonstrated lower chances of receiving sufficient water supply. Moreover, if the councillor is affiliated to the AITC then the chances of getting sufficient water increases by more than 27%. Religion plays an important role in the reception of sufficient water supply. The chances of sufficient water supply were reduced by at least 16% for Muslim households and by at least 26% for wards represented by a Muslim councillor. We have used the number of family members as a control variable. An increase in the number of family members reduces the chances of sufficiency of water. The variable, per capita monthly consumption expenditure, acts as a proxy for income level of households. The richer the household, the lower the chances of satisfaction with the availability of water. This may be due to the influence of response bias. Overall, political competition has a negative effect, political fragmentation has a positive but insignificant effect, while the combination of high fragmentation and high competition has a negative effect on service delivery.
6.2 Quality of water The five models illustrated in Table 9 present the maximum likelihood estimates of determinants of good quality water supply (‘good_water’). The coefficient of PL, GL, MUSLIM, FEED, CONTEST, AITCC, VOTE, MARGIN, CMARGIN and HFLM are positive while the coefficients of NN, CK, SCST and MC are negative. All independent variables are statistically significant. MUSLIM and FEED are significant in all four models. SCST is also significant in three models. All these five models have been able to achieve around 80% correct prediction. Quality of water supply greatly depends on the legal status of the slum area. The probability of getting good quality water increases by 40% if slums are situated on private land, but the probability falls by the same extent if slums are located in non-notified areas. The chances of good quality water increase by 15% in slums located on government land. Caste and religion significantly affect access to good © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
269
quality water. The probability of good quality slides down by up to 21% for households belonging to poorer castes. Whereas the chances of getting good quality water increase for Muslim households. Response bias of the Muslim community, which suffers from acute water shortages and lack of education, could be one of the reasons for expressing good quality of water. However, this response bias gets
Table 9: Determinants of availability of good quality water Dependent variable: good quality (good_quality)
Model I
Non-notified Area (NN)
0.40#*** (24.87)
Private Land (PL)
Model II
Model III
Muslim Household (MUSLIM) Less developed Caste Household (SCST) Feedback to Local Government (FEED) Election fought per 10,000 electorate (CONTEST) Muslim Councillor (MC)
0.15* (2.76) 0.18*** (12.07) 0.20*** (18.61) 0.21*** (7.71) 0.15*** (12.99)
0.18*** (12.07) 0.20*** 0.39*** (18.61) (47.7) 0.21*** (7.71) 0.15*** 0.23*** (12.99) (24.7) 0.12*** (27.28) 0.60*** (57.64)
AITC Councillor (AITCC) % vote obtained (VOTE)
0.76*** (93.70) 0.27*** 0.21*** (19.89) (19.62) 0.19** 0.13* (5.82) (2.82) 0.20*** 0.13** (11.7) (6.40) 0.12*** (24.16) 0.22*** (7.88) 0.32*** (16.65)
2.19*** (49.95) 0.61*** (14.13)
Re-elected Councillor (REELEC) Margin (MARGIN) Change of Margin (CMARGIN) High Fragmentation & Low Margin (HFLM) Percentage Correct LR chi2 Prob > chi2 Pseudo R2 No of observations
Model V
0.40*** (24.87)
Government Land (GL) Central Kolkata (CK)
Model IV
80.1 94.63 0 0.1535 513
80.1 94.63 0 0.1535 513
81.7 196.93 0 0.3195 513
81.4 191.71 0 0.311 513
1.44*** (31.25) 0.41*** (7.45) 0.13* (3.13) 78.9 130.90 0 0.21 513
Note: Numbers in the parentheses are Wald chi-square statistics. Marginal Effects, not the coefficients. *** indicates significant at 1% level, ** at 5% level and * at 10% level. © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
270
Indranil De and Tirthankar Nag
diluted once we consider the other communities in a jurisdiction. The chances of good quality water fall by up to 60% where the local councillor is Muslim. Furthermore, the probability of getting good quality water is lower by at least 18% in Central Kolkata slums that are Muslim dominated. Local government appears to play a major role in maintaining the quality of water supply. Quality of water improves if the current councillor happens to have won elections with a higher percentage of votes and a higher and improved margin as compared to the preceding elections; that is to say, if political competition reduces. The higher the political fragmentation (measured by the number of contestants per ten thousand voters), the better the chances of good quality water. Wards characterised by both high competition and high fragmentation have higher chances of receiving good quality of water supply. The probability of getting good quality water increases by 32% if the councillor is affiliated to the AITC. The chances of getting good quality water increases by up to 23% if the local government is made aware of the feedback of households on quality of water. Hence, organic linkage with local government improves services. Furthermore, if the local councillor is a repeat (re-elected) councillor of the previous municipal council, then the chances of getting good quality water fall by 61%. This could be due to less effort on the part of the councillors in governance. Overall, the results illustrate the fact that political competition has a negative impact, political fragmentation has a positive impact, and the combination of political competition and fragmentation has a positive impact on the quality of water.
6.3 Drainage The maximum likelihood estimates of the determinants of ‘good_drainage’ are illustrated in Table 10. The coefficients of NK, CK, SCST, MC, and KEIP are positive while the coefficients of PL, CONTEST, AITCC, VOTE, MARGIN, CMARGIN and HFLM are negative. All independent variables are statistically significant. SCST and MC are significant in all the models and hence robust. With respect to the overall model fit, all three Tobit models appear to be statistically significant. Slums located on private land are less likely to get services like cleaning of drainage within a fortnight. Households from North and Central Kolkata slums stand good chances of drainage cleaning within a fortnight. This could be due to high population density in these regions. Poorer caste households have higher chances of receiving good drainage services. Moreover, households represented by a Muslim councillor have higher chances of good drainage services. Candidates victorious in more highly competitive elections (lower percentage of votes, less margin and reduced margin as compared to previous elections) demonstrate better chances of good drainage service delivery. However, political fragmentation (measured by number of contestants per ten thousand voters) reduces the chances of better service delivery. Moreover, high political competition coupled with high political fragmentation is likely to lead to lower delivery of services. Councillors not associated with the majority party might try to showcase their work by delivering services such as drainage, which is under their direct control. The Kolkata © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Change of Margin (CMARGIN)
Margin (MARGIN)
KEIP Implemented (KEIP)
% vote obtained (VOTE)
AITC Councillor (AITCC)
Less developed Caste Household (SCST) Election fought per electorate (CONTEST) Muslim Councillor (MC)
Per capita consumption expenditure (MPCE) Muslim Household (MUSLIM)
Central Kolkata (CK)
North Kolkata (NK)
Private Land (PL)
Dependent variable: good drainage (good_drainage)
4.18*** (252.49)
0.25*** (11.83) 0.09*** (15.6) 0.83*** (270.27)
0.39*** (31.02)
Model I (coeff.)
3.06*** (247.12)
0.54*** (182.25)
0.11** (5.48)
1.99*** (469.59)
Model II (coeff.)
0.3*** (18.4) 0.07*** (10.05) 0.22*** (7.29) 0.60*** (63.04) 3.01*** (67.73)
0.48*** (39.31)
Model III (coeff.)
1.33*** (8.47) 0.55*** (121.22) 2.90*** (83.17)
0.72*** (151.29)
Model IV (coeff.)
Table 10: Determinants of good drainage service
0.43*** (57.15) 0.37*** (20.07) 6.29*** (170.56) 0.68*** (100.40)
0.15*** (5.29)
Model V (coeff.)
0.69*** (33.76)
5.20*** (399.20) 0.73*** (107.33)
0.10*** (24.80) 0.52*** (54.17)
0.34*** (10.69) 0.0001** (5.66)
Model VI (coeff.)
Dangers of decentralisation in urban slums
271
2.52*** (245.24) 373.89 0 0.3192 541
Model I (coeff.)
3.63*** (676) 920.67 0 0.7861 541
Model II (coeff.)
2.37*** (183.06) 423.78 0 0.3618 541
Model III (coeff.)
1.21*** (40.32) 477.65 0 0.41 540
Model IV (coeff.)
0.81*** (69.22) 3.62*** (274.23) 482.27 0 0.41 540
Model V (coeff.)
Note: Numbers in the parentheses are Wald chi-square statistics. *** indicates significant at 1% level, ** at 5% level and * at 10% level.
LR chi2 Prob > chi2 Pseudo R2 No. of Observations
High Fragmentation & Low Margin (HFLM) constant
Dependent variable: good drainage (good_drainage)
Table 10: Continued
0.39 536
2.82*** (267) 456.86
Model VI (coeff.)
272 Indranil De and Tirthankar Nag
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
273
Environmental Improvement Project (KEIP) has demonstrated a positive impact on drainage services. Overall, the results suggest that political competition has a positive impact, political fragmentation is negative, while a combination of high competition and high fragmentation has a negative impact on drainage services.
7
Conclusion
The theoretical rationale posited in the study suggests that the higher the level of local government from which a service is managed and the more varied the awareness regarding the service, the lower the chance of being affected by clientelism. Treatment of water is done at higher level, and quality awareness varies across individuals, therefore clientelism does not affect quality of water. The lower the level of local government from which a service is managed and the higher the measurability of services standards, the more the service is susceptible to clientelism. Distribution of water (quantity) and drainage service (number of times drains cleaned) is managed from a much lower level of local government and services are measurable. Therefore, these are more likely to be affected by clientelism than water quality. Following the criteria spelt out by Prud’homme (1994), drainage service is more suitable for decentralisation, but has a higher chance of underprovision due to clientelism. Therefore, awareness and measurability of service should be taken as criteria to scrutinise the suitability of service for decentralisation. The higher the variation of awareness and the lower the measurability of service standards, the less the service is susceptible to clientelism. Political competition results in a lower provision of services deemed to be important and resource intensive. Water supply (both quantity and quality), being an important and resource intensive service, is less provided in wards where political competition is high (councillors do not have substantial political leverage over others) and councillors are not affiliated to majority party of the state and corporation. However, delivery of less important and resource intensive services (such as drainage) improves with increasing political competition and the affiliation of councillors to political parties other than the majority party. The social and religious identities of households and councillors have a significant impact on service provision. Less developed caste and Muslim households, and wards represented by Muslim councillors, are likely to receive a lesser quantity of water but more drainage service. The higher importance and resource intensiveness of water supply service compared to drainage service leads to capture. This study suggests that the importance of the service and its resource intensiveness should be considered as additional criteria, along with awareness and measurability, in order to examine the suitability of a service for decentralisation. The article highlights the measurability, monitoring and awareness of public services to deal with clientelism and capture at the local level. The negative implication on equity due to decentralisation can only be countered through more methodical intervention at local level in the domain of public service delivery, rather than through a deconcentration-oriented decentralisation process. The intervention should be oriented towards people’s participation in local governance, which has © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
274
Indranil De and Tirthankar Nag
only been hypothetical hitherto. The central and state level initiatives over legal and constitutional empowerment of local bodies have not been able to serve the poor and ethnic minorities due to the inefficiencies and malfunctioning of local bodies. Therefore, broader reforms for strengthening local governments are needed, but at the same time it is important to provide financial and human resources to local governments, to monitor and benchmark services at local level and to generate awareness about better public services. first submitted February 2014 final revision accepted October 2015
References Aijaz, R (2010) Water for Indian Cities: Government practices and policy concerns. ORF Issue Brief No.25. New Delhi: Observer Research Foundation. Asthana, A. N. (2003) ‘Decentralization and Supply Efficiency: The case of rural water supply in Central India’, The Journal of Development Studies 39(4): 148–59. Bardhan, P. and Mookherjee, D. (2000) ‘Capture and Governance at Local and National Level’, American Economic Review 90(2): 135–9. Bardhan, P. (2002) ‘Decentralisation of Governance and Development’, Journal of Economic Perspectives 16(4): 185–206. Benit-Gbaffou, C. (2011) ‘“Up Close and Personal”: How does local democracy help the poor access the state? Stories of accountability and clientelism in Johannesburg’, Journal of Asian and African Studies 46(5): 453–64. Besley, T. and Burgess, R. (2002) ‘The Political Economy of Government Responsiveness: Theory and evidence from India’, The Quarterly Journal of Economics 117(4): 1415–51. Canel, E. (2001) ‘Municipal Decentralization and Participatory Democracy: Building a new mode of urban politics in Montevideo City?’, European Review of Latin American and Caribbean Studies 71: 25–46. Chandra, M. (2004) Bridging Everyday and High Politics: The 74th CAA and Inclusion in Kolkata, India. European Association of South Asian Studies Paper 31. Lund: Lund University Chanie, P. (2007) ‘Clientelism and Ethiopia’s Post-1991 Decentralisation’, The Journal of Modern African Studies 45(3): 355–84. Chattopadhyay, S. (2012a) ‘Decentralisation and Delivery of Urban Basic Services: The West Bengal experience’, Development in Practice 22(1): 57–70. Chattopadhyay, S. (2012b) ‘Decentralized Urban Governance Participation and Accountability in West Bengal’s Municipalities’, Environment and Urbanization Asia 3(1): 185–202. Contractor, Q. (2012) ‘Quest for Water: Muslims at Mumbai’s periphery’, Economic and Political Weekly 47(29).
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
Dangers of decentralisation in urban slums
275
Das, K. (2009) Agency and Access under Decentralized Governance: Water supply and sanitation in Kolkata City. GIDR Working Paper Series No.191. Ahmedabad: Gujarat Institute of Development Research. Garcıa-Guadilla, M. P. and Perez, C. (2002) ‘Democracy, Decentralization, and Clientelism: New relationships and old practices’, Latin American Perspectives 29(5): 90–109. Gay, R. (1990) ‘Community Organization and Clientelist Politics in Contemporary Brazil: A case study from suburban Rio de Janeiro’, International Journal of Urban and Regional Research 14(4): 648–66. Heredia, B. (1997) Clientelism in Flux: Democratization and interest intermediation in contemporary Mexico. Working Paper. Mexico: CIDE (Centro de Investigaci on y Docencia Econ omicas). Holbrook, T. M. and Van Dunk, E. (1993) ‘Electoral Competition in the American States’, American Political Science Review 87(4): 955–62. IDFC (Infrastructure Development Finance Company) (2011) Indian Infrastructure Report 2011. New Delhi: Oxford University Press Keefer, P. (2002) Clientelism, Credibility and Democracy. Washington, DC: World Bank, Development Research Group. Keefer, P., Narayan, A. and Vishwanath, T. (2003) The Political Economy of Decentralization in Pakistan. Washington, DC: World Bank, Development Research Group. Keefer, P. and Khemani, S. (2004) ‘Why Do the Poor Receive Poor Services?’, Economic and Political Weekly 39(9). Kitschelt, H. and Wilkinson, S. I. (2007) ‘Citizen–Politician Linkages: An introduction’, in H. Kitschelt and S. I. Wilkinson (eds), Patrons, Clients, and Policies: Patterns of democratic accountability and political competition. Cambridge: Cambridge University Press. Lake, D. A. and Baum, M. A. (2001) ‘The Invisible Hand of Democracy: Political control and the provision of public services’, Comparative Political Studies 34(6): 587–621. Lazar, S. (2004) ‘Personalist Politics, Clientelism and Citizenship: Local elections in El Alto, Bolivia’, Bulletin of Latin American Research 23(2): 228–43. Lemarchand, R. (1972) ‘Political Clientelism and Ethnicity in Tropical Africa: Competing solidarities in nation-building’, American Political Science Review 66(1): 68–90. Litvack, J., Ahmad, J. and Bird, R. (1998) Rethinking Decentralization in Developing Countries. Sector Study Series. Washington, DC: World Bank. Makumbe, J. (1996) Participatory Development: The case of Zimbabwe. Harare: University of Zimbabwe Publications. McKenzie, D. and Ray, I. (2009) ‘Urban Water Supply in India: Status, reform options and possible lessons’, Water Policy 11(4): 442–60. Misra, S. and Goldar, B. (2008) ‘Likely Impact of Reforming Water Supply and Sewage Services in Delhi’, Economic and Political Weekly 43(41): 57–66. Mtisi, S. and Nicol, A. (2003) Water Points and Water Policies: Decentralisation and community management in Sangwe Communal Area, Zimbabwe. Sustainable © The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
276
Indranil De and Tirthankar Nag
Livelihood in Southern Africa Research Paper 15. Brighton: Institute of Development Studies, University of Sussex. Oliver, P., Marwell, G. and Teixeira, R. (1985) ‘A Theory of Critical Mass: I) Interdependence, group heterogeneity, and the production of collective action’, American Journal of Sociology 91(3): 522–56. Ostrom, E. (2000) ‘Collective Action and the Evolution of Social Norms’, Journal of Economic Perspectives 14(3): 137–58. Pradhan, K, B. and Roy, P. K. (2003) The Well Being of Indian Households. Noida, India: Tata Mcgraw-hill Publishing Company. Prud’homme, R. (1994) On the Dangers of Decentralization. Policy Research Working Paper 1252. Washington, DC: World Bank. Ranganathan, M., Kamath, L. and Baindur, V. (2009) ‘Piped Water Supply to Greater Bangalore: Putting the cart before the horse?’, Economic and Political Weekly 44(33). Skilling, D. and Zeckhauser, R. J. (2002) ‘Political Competition and Debt Trajectories in Japan and the OECD’, Japan and the World Economy 14(2): 121–35. Slaymaker, T. and Newborne, P. (2004) Implication of Water Supply and Sanitation Programmes under PRSPs. London: Overseas Development Institute, Water Policy Programme. Srinivasan, K. and Mohanty, S. K. (2004) ‘Deprivation of Basic Amenities by Caste and Religion’, Economic and Political Weekly 39(7). WSP (Water and Sanitation Programme) (2004) ‘Rural Water Supply and Sanitation in Africa: Global learning process on scaling up poverty reduction’, Shanghai Conference, 25–27 May. WSP (2003) Governance and Financing of Water Supply and Sanitation in Ethiopia, Kenya and South Africa: A cross country synthesis. Sector Finance Working Paper No. 5. New York, NY: United Nations Development Programme, Water and Sanitation Programme.
© The Authors 2016. Development Policy Review © 2016 Overseas Development Institute. Development Policy Review 34 (2)
