Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches

Utilities Policy xxx (2017) 1e13

Contents lists available at ScienceDirect

Utilities Policy journal homepage: www.elsevier.com/locate/jup

Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches Sriharini Narayanan, A. Thillai Rajan*, Paul Jebaraj, M.S. Elayaraja Indian Institute of Technology, Madras, India

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 June 2015 Received in revised form 10 January 2017 Accepted 10 January 2017 Available online xxx

In the provision of basic infrastructure services to the urban poor, limited rigorous evidence on the most effective service delivery approaches is available. This meta-analysis synthesises the evidence on the effectiveness of bottom-up approaches that is characterized by the strong involvement of alternate service providers such as NGO's and CBO's in improving access to electricity, water supply, and sanitation services for the urban poor. Although bottom-up approaches are espoused, we find that they do not have a statistically significant effect. This trend was consistent for all dimensions of access: connectivity, affordability, adequacy, and effort and time. However, our findings also show that bottom-up approaches may be more effective in the water and sanitation sectors than in the electricity sector. When bottom-up approaches involve active participation from the community, the results are significantly positive. Our study suggests that innovations to bottom-up approaches that facilitate active community participation can be an effective way to increase access to basic services among the urban poor. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Basic infrastructure services Urban slums Participation Tenure security

1. Introduction Providing access to basic infrastructure services such as water, sanitation, and electricity to the urban poor in Low and Middle Income Countries (LMICs) poses a major challenge to policy makers. The overall progress made on achieving the targets set for providing basic services to slums has not been enough to match the expansion of informal settlements in developing countries. Close to 828 million or 33 percent of the urban population in developing countries reside in slums without access to basic services (UNHabitat, 2012). Though the international community had responded to the challenge with an increased commitment for Official Development Assistance (ODA) to the poorest countries (OECD, 2013), much remained to be achieved. In sanitation, for example, 2.5 billion people still lacked access to toilets and latrines (MDG, 2013). Unprecedented growth in slums, poor urban planning, and supply-side challenges widened the gap between the demand and supply of basic infrastructure services (Bakker, 2007). Perspectives on access to basic services have expanded over time from state-led initiatives to include market-based, private sector-oriented, and user-participative approaches (Dagdeviren

* Corresponding author. E-mail address: [email protected] (A.T. Rajan).

and Robertson, 2011). Traditionally, government agencies were vested with responsibility for providing universal access to services such as water, sanitation, and electricity. However, governments lacked the financial resources, the institutional capacity, and often the political will required to extend coverage of basic services to the slum settlements (Brooke and Smith, 2001). The state-owned utilities, which were the main providers of infrastructure services in developing countries, were characterized by high costs and poor performance. In order to increase efficiency and improve performance, several governments implemented a program of reform and restructuring of their utilities from the mid-1990's (Kessides, 2004; Clifton et al., 2011). While these reforms may have served to bring about overall fiscal discipline and managerial efficiency, they often did not include an institutional mechanism to reinforce the responsibility of the state to provide basic services to slums and low-income groups (Brooke and Smith, 2001). This had resulted in the emergence of alternate service providers led by non-government organizations (NGOs), community-based organizations (CBOs), and other small-scale service providers to fill the gap in basic infrastructure services in urban slums (Nijman, 2008; Ibem, 2009). In tandem, some state agencies, having understood their limitations in being able to service slums, started orchestrating the involvement of these alternate service providers (Dagdeviren and Robertson, 2011; Nijman, 2008; Brooke and Smith, 2001;

http://dx.doi.org/10.1016/j.jup.2017.01.002 0957-1787/© 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

2

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

UNCTAD, 2008; and Baud and Dhanalakshmi, 2006). Formal partnership arrangements among governments, communities, and service providers have been seen as a practical solution to address the problems of access to basic services in slums (Ibem, 2009). 1.1. Dimensions of service delivery Three main players are involved in the provision of basic services to the urban poor: the state, the private sector, and the voluntary sector. The extent of involvement of each of the players in the slums is influenced by economic, social, and political factors. Although the government once was the main provider of basic services in several countries, a brief look into the history of reforms suggests that in recent years various forms of private sector participation has been put in place (Kessides, 2004; Clifton et al., 2011). Whilst there have been cases of successful infrastructure reform through privatisation, under-investment in infrastructure in developing countries remains a concern (UNCTAD, 2008; Dagdeviren and Robertson, 2011). Devkar et al. (2013) in their systematic review found that privatisation has had varying effects on access to basic services by the poor. They claim that the involvement of private sector does not improve access to the poor unless specifically supported with targeted investment programs and different forms of assistance from the public sector. Sector-wise differences in impact are also relevant; reforms in the electricity sector have led to expanded coverage across the urban populace, including the urban poor (Brooke and Smith, 2001) whereas the water and sanitation sectors remain neglected (Otiso, 2003). The inadequate provision of infrastructure by the government has led to the emergence of small-scale providers who attempt to fill the infrastructure gap based on local knowledge and innovative
n and McGranahan, 2006). Outside of the business practices (Kjelle formal utility system in developing countries, several private, small-scale service providers are actively involved in filling the infrastructure service gap (Nijman, 2008). Small-scale providers include water vendors, local community groups, and NGOs. Their role in ensuring access to basic services has gained prominence in recent years. Multilateral institutions such as the World Bank, UN-Habitat, and DFID have recognised the need to include the small-scale suppliers in the supply chain by providing them with formal recognition as well as regulatory and policy ad
n and McGrajustments (Mundial, 2004; UN-Habitat, 2003; Kjelle nahan., 2006). UN-Habitat went a step further to recommend community-based water-supply schemes for low-income peri-urban areas (UN-Habitat, 2003). Strengths attributed to these smallscale providers include the local knowledge that helped to tailormake service delivery in slums facing constraints such as tenure security, poor layouts, and low affordability (USAID, 2004; Burra et al., 2003; Weitz and Franceys, 2002). Although small-scale operators are recognised as important service providers in informal settlements, they often function illegally and take considerable risk in providing services to the urban poor (Dagdeviren and Robertson, 2011). The nature of service and the infrastructure network required for service delivery also determines the type of service provider and the service delivery approach in slums. For instance, the electricity sector exemplifies of a service that required centralised planning and implementation (Manzetti and Rufin, 2006; Scott et al., 2005, Shrestha et al., 2008). While provision of water follows a networked approach similar to that of electricity, government agencies have only been partially successful in improving access to water in slums (Hossain, 2012; Connors, 2005; Ghafur, 2000). Some of the pitfalls of the top-down approach in water sector are attributed to the lack of stimuli to make progress in addressing poor service delivery, lack of legal mandates and financial resources for network

expansion in informal slums, and the lack of knowledge required to overcome the challenges of cost-recovery in slums (Hossain, 2012; Connors, 2005; Ghafur, 2000). Water supply also lends itself to the implementation of decentralized solutions, such as wells. Sanitation in slums is even worse when compared to that of electricity and water due to the failure of government service providers in terms of planning, design and construction, and the lack of maintenance (Bapat and Agarwal, 2003; Burra et al., 2003; Hobson, 2000). Few authors conclude that the top-down service delivery in sanitation has resulted in the construction of community or public toilets that are neither adequate nor well maintained in most developing countries (Ghafur, 2000; Roma and Jeffrey, 2011). Apart from infrastructure requirements, an important aspect that affected service delivery is the legal status of slum settlements. Scott et al. (2005) state that the extent to which urban slums are provided connections hinged on the legal status of the slum settlement. Informal slums that had no legal status have lesser access to services when compared to formal slums that enjoy de jure tenure security (Chandrasekhar, 2005). A factor that distinguishes the government providers from the small-scale providers is the extent of the involvement of the slum community and its inclusivity in formulating specific strategies to improve living conditions. There exists a spectrum of approaches that not only involve varying levels of inclusivity but also different forms of organisation that participate in the planning and delivery of services. Participation by different stakeholders, such as community residents, government officials, NGOs and CBOs, determine whether the approach is top-down or bottom-up. Literature on slum up gradation and slum resettlements show that providers in none of the three service sector can single-handedly meet the growing needs of the urban poor (Otiso, 2003). Therefore, there is an urgent need to understand the results of the different approaches to service delivery in urban slums in order to assess the efficacy of the various interventions implemented so far and to design appropriate interventions and policy frameworks based on these observations (Brooke and Smith, 2001).

1.2. Focus and structure of the paper The gap in evidence-based research in basic services delivery in urban slums has also impacted policy-making. Despite several initiatives, there is limited evidence regarding the best choice of context-specific delivery models for programming (DFID, 2012). UNCHS (2006) also noted that there are very few studies that demonstrate the extent to which local communities have been involved in the provision of infrastructure in slums (UNCHS, 2006). Although studies such as the one done by Bel et al. (2010) use metaregression analysis to provide empirical evidence on the impact of privatisation on the cost of local municipal services, such as water supply and solid waste management, the study provides limited evidence on the specific factors and context that improve access. The objective of this paper is to analyse the effectiveness of involving alternate service providers in the provision of basic infrastructure services for slums and low-income settlements. While various city or region specific studies have analysed the impact of alternate service providers, a holistic perspective on the effectiveness of alternate service providers has been missing. We attempt to provide such a perspective by synthesizing the evidence from different studies using meta-analysis techniques. The rest of this article is organized as follows. In Section 2 we describe the key concepts, methodology, and data. Section 3 presents the results of the meta-analysis and discusses the findings. Summary and implications for policy are discussed in Section 4.

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

2. Methods 2.1. Conceptual overview Broadly, the study synthesized the evidence from Low- and Middle-Income Countries (LMICs) due to the severity of the problem of lack of access to basic services. In terms of domain, the study focused on the water, sanitation, and electricity sectors. We chose to focus on these sectors because of their impact on improving the quality of daily lives. Although all the three services have been grouped together as basic services, each of them is unique as a sector and has been dealt with separately. Approaches to deliver basic services have been classified into two approaches: (1) top-down and (2) bottom-up. The traditional top-down approach was characterized by centralised planning and government provision of services, with little or marginal participation of other stakeholders. On the other hand the bottom-up approach is characterized by strong involvement from social welfare organizations such as NGO's and CBO's (the alternate service providers) in the planning, implementation and delivery of services. In reality, there may be very few pure top-down or bottomup approaches since the majority of the programs fall in between these two poles, with varying degrees of decentralisation. However, for the purpose of this study, we have considered the top-down approach as the default service provision approach in slums against which the effectiveness of bottom-up interventions was compared. The outcome variable of interest for this study is the level of access to water, sanitation, and electricity services since access is an integral element in achieving the Millennium Development Goals (MDGs, Global Urban Observatory, 2003). While the traditional definitions of access would mean connectivity, we broadened the indicators to include other dimensions such as affordability, adequacy, and effort and time (UN-Habitat, 2002). The definitions adopted for each of these dimensions are given below:
Connectivity: Proportion (or mean number) of households with access to water supply (public tap or individual in-house tap), sanitation facilities (individual toilet or community toilet) or legal electricity connections at the household level.
Affordability: The proportion of households who found the services affordable (or) the mean cost incurred in accessing the services for the dwellers.
Adequacy: This dimension measured the proportion of households who found the services adequate (or) the mean level of consumption of services.
Effort and time: Proportion of households who were able to access the services without excessive effort and time (or) the distance travelled for accessing the services (mainly relevant for water and sanitation). Since outcome variables (level of access) are impacted by various factors such as the sector, the type of facility, and the context of the operating environment, we have also accounted for these variables in our analysis (adapted from Pettigrew, 1987; Dawson, 1994; Nelson and Dowling, 1998). Specifically, the analysis was conducted separately for each dimension of access within each sector and for public and household connections. Under the context of operating environment, we considered three factors: geographical region, type of settlement, and extent of community participation. When adequate evidence was available, we separately analysed the results by geographical region. Our data were from the following three broad LMIC regions: Asia, Africa, and South America. Type of slum settlement was classified into one of the following types: formal, informal, and low-income. Formal

3

slums are those that are recognised by the state and where dwellers have either de jure or de facto security of tenure. Informal slums are those that are considered as illegal settlements and dwellers in such slums do not have any security of tenure. Poor households not located in a defined slum (either formal or informal) but distributed across the city in non-slum areas were classified as part of a lowincome settlement. We considered the level of participation from the community to account for gradations in the degree of inclusivity. Participatory planning was understood to mean the involvement of different stakeholders such as community residents, officials from government and other institutions like NGOs and CBOs in the planning and implementation of a service. Possible forms of community participation (adapted from UN-Habitat, 2003; Kingston, 1998) are: passive participation, participation through giving, participation through consultation, participation through contribution, participation through partnership, and participation through selfmobilisation. Broadly, for the purpose of this study, we considered two categories of participation: interventions where there was no active participation (i.e., passive participation) and interventions where there was some level of participation from the community such as consultation, contribution, or self-mobilisation. 2.2. Methodology The methodology for this study broadly consisted of the following steps. First, those studies that qualified for inclusion in the review were identified. Second, the evidence from those studies was synthesized using appropriate meta-analysis techniques. In this section, we briefly describe these two steps. 2.2.1. Study identification methods Study identification consisted of the following steps (1) identifying inclusion and exclusion criteria (refer appendix 1), (2) identify the sources from which studies could be obtained, and (3) search strategy. Inclusion and exclusion criteria essentially helped in setting the boundaries of the research from which the sources and search strategy follow. The different sources that were targeted for identifying the studies for this review were: (1) electronic journal databases such as ScienceDirect, EBSCO, and ProQuest; (2) the websites of IADB, World Bank, UN-Habitat, City Alliance, UNDP, ADB, DFID, EPPI, Practical Action, CLIFF, AusAid, Slum Upgrading Facility, Cochrane Systematic Review, Campbell Systematic Review, Research for Development, Energy Sector Management Assistance Program, Global Network on Energy for Sustainable Development, WaterAid, and 3ie; (4) references from studies that qualified for inclusion in the review; and (4) directly corresponding with the authors who published frequently on topics relevant to this review. The search strategy also followed a systematic process that started with the hand search of journals followed by website search, search of electronic databases, reference search and direct correspondence with key authors identified through the hand search. Hand search and electronic database search yielded 21,919 studies whereas search of websites, reference search, and author correspondence yielded 3580 studies. All these studies were screened first by applying the exclusion criteria; those that were excluded were not evaluated further. The studies that remained were then evaluated on the basis of the inclusion criteria. Only those studies that met all the inclusion criteria were considered for the next stage in the multi-stage screening and selection process. EPPI-Reviewer software was used to manage the entire search process and remove duplications. The citations, including the abstracts and the document in PDF format were imported into the Reviewer for screening. In some cases, where the study details

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

4

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

could not be directly imported to the EPPI-Reviewer, it was entered manually so that the entire repository of studies could be managed in the EPPI-Reviewer software. 2.2.2. Screening and selection process The different steps in screening process consisted of quick screening of the title followed by screening of the abstract and then a full text screening of those papers that cleared the first two stages. The articles that qualified for inclusion after screening and selection process were assessed for quality using a critical appraisal tool. Annamalai et al. (2016, p. 165e168) provides the appraisal tool used in this review. Each study was appraised for quality by two members of the review team, working independently. Papers that obtained a rating of more than 60 out of a possible 72 were considered as high quality studies, score ranging between 35 and 60 as medium quality and lower than 35 as low quality studies. The cut off scores were quality were arrived after extensive discussions within the research group. 2.3. Data description Twenty-one studies that fulfilled the inclusion and exclusion criteria were included for meta-analysis after the screening and selection process (refer Appendix 2). These 21 studies provided 71 findings of evidence that were used for analysis. Multiple findings of evidence could be obtained from several studies, since many studies covered different sectors, populations, and access outcomes. However, if there was more than one finding of evidence on an outcome using the same dataset, only the most robust evidence was taken for the analysis. Avoiding multiple findings of evidence for the same outcome using the same dataset ensured the independence of observations. Table 1 provides a classification of the evidence on different parameters. Panel A gives the details on the evidence for different dimensions of access. The highest amount of evidence was reported for connectivity, followed by affordability. Since the findings of evidence for affordability, adequacy, and effort and time were 14, 4, and 4, respectively, detailed sub-analysis for various contextual

environment factors could not be done for these outcomes. However, for connectivity, we could perform sub-analysis for various contextual factors such as geographical region, type of slum, and extent of community participation. Panel B gives the evidence for different sectors. The amount of evidence for the water sector was the highest, followed by that for sanitation. These numbers can also be seen as an indication of the importance of these services for slums and low-income population, the challenges that exist in provisioning these services, and the consequent research interest. Electricity has received the least priority probably because access to illegal electricity is quite common in urban slums and there are alternate energy fuels that can substitute for electricity. This is perhaps reflected in the limited number of research publications available in this sector. Panel C gives the evidence by region. Though the evidence in Asia are the highest, we have reasonable findings of evidence for Africa and South America regions also. Panel D gives the evidence on connectivity in the different types of slums. The amount of evidence in informal slums is the highest at 38 followed by formal slums at 6 and low-income settlements at 5. These results show a high research interest in tenure security and the study of informal slums. Panel E provides the evidence on connectivity when there is some form of participation by the community. Although we looked at different types of participation, for the purpose of analysis, the evidence was grouped into limited or no participation and active participation. The results show reasonably high evidence for participation, which is an encouraging finding for the purposes of our study. In terms of the quality of the evidence, 20 studies were rated either high or medium and only one study was rated low. We thus consider the overall evidence base to be not just representative but also of good quality. Please refer to appendix 3 for the quality appraisal results.

2.4. Meta-analysis Meta-analysis, is a systematic approach to synthesize quantitative evidence from results of previously published studies in order to arrive at conclusions on a body of research (Haidich, 2010;

Table 1 Classification of evidence used for meta-analysis. Bottom e up approach

Panel A: Dimensions of Access 1 2 3 4 Total Panel 1 2 3 Total Panel 1 2 3 4 Total Panel 1 2 3 Total Panel 1 2 Total

Connectivity Affordability Adequacy Effort and time

49 14 4 4 71

Water Sanitation Electricity

45 21 5 71

Asia Africa South America Central and North America

36 4 27 4 71

Formal Slum Informal Slum Low-income settlements

6 38 5 49

Limited or no participation Active community participation

33 16 49

B: Sectors

C: Region

D: Type of Slum - Connectivity

E: Community Participation - Connectivity

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

Donna et al., 2000). Meta-analysis combines a collection of evidence from independent studies on a particular topic to evaluate its magnitude and statistical significance of the summary effect. In this way, it plays a major role in effectiveness research by providing coherent results with statistical significance to identify those interventions that have a strong impact on outcomes. An advantage of meta-analysis is that it helps in a more organized and quantifiable review of the literature, while providing a defined precision and objectivity of the summary estimate (Rosenthal, 1991). However, overall findings are less conclusive when the included studies suggest opposing effects. In order to estimate the extent of such opposing effects, statistical test for heterogeneity was done to assess the variation across studies (Higgins et al., 2003). In order to be able to make appropriate inferences from our findings, we ensured that all studies included in a single metaanalysis had outcomes measured in the same way. When results were measured differently for the same outcome, a separate metaanalysis was conducted. While the proportion of people having connections was the single unit of measurement for connectivity across all studies, it was not the case for other access indicators such as affordability, adequacy, and effort and time. We had outcomes measured in proportions (i.e., the proportion of people who found the services affordable) or in mean values (i.e., cost incurred by households for unit consumption of the service, or distance travelled to access the service). In addition to the evidence on outcome, the dataset also included other characteristics of the study, such as the sample size, sector (water, sanitation or electricity), type of access (individual or community), level of participation (some level of participation or no participation), type of service delivery (top-down or bottomup), regional location of the slum (Asia, Africa, South America, or Central and North America), and type of slum (informal, formal or low-income). These study characteristics were used to separate the evidence for meta-analysis for various sub-group analysis. The resultant dataset was analysed using STATA. Since most of the reported findings were in proportions, calculating an odds-ratio was chosen as the most appropriate measure to synthesize the results. When studies reported evidence in mean and standard deviation they were converted to standardized mean difference (SMD) by calculating the mean difference. Finally, to have a unique measure for performing meta-analysis, we transformed all the findings into a common metric, viz., the log odds ratio. SMD was converted into log odds ratio as given below based on the formula suggested by Borenstein et al. (2011).

LogOddsRatio ¼ SMD

p √3

(1)

The variance for log odds ratio was calculated as follows:

VLogOddsRatio ¼ VSMD

p2 3

(2)

Variance was converted to standard error of the log odds ratio for analysis. Using relevant commands in STATA to conduct meta-analysis, we calculated the pooled effect sizes (ES) as a measure of odds ratio by analyzing log odds ratio and its standard error. To ensure that the pooled ES were appropriate, we conducted separate subanalysis for evidence that can be appropriately pooled together. We hypothesized that if the effect size was significantly different from 1, there was a significant difference in the level of access between top-down and bottom-up approaches. We used the random effects model in our analysis because of the heterogeneity across the included studies, and the true effect size also varied from study to study. We summarized pooled effect sizes based on the sub-

5

analysis using random effect models with DerSimonian and Laird Method to identify the between-study variance. Heterogeneity was analysed from the I2 values as it presented the observed dispersion in proportions. Heterogeneity was inevitable in our review due to the inclusion of studies with sampling variation of small and larger studies, diverse study populations, different study methodology, and different methods used for collecting primary data. Studies in the review may not be systematically representative of the population. In meta-analysis, if the sample of studies included for synthesis is biased, then the validity of the results is affected (Rothstein et al., 2006; Egger et al., 2000). Funnel plot asymmetry test was used to examine whether the association between intervention effect size (log odds ratio) and study size (standard error) measure is greater than might be expected to occur by chance (Sterne et al., 2000). We used Egger's test (Egger et al., 1997) and Begg's test (Begg and Mazumdar, 1994) to assess bias. While the former uses linear regression but is based on the efficient score and its variance, the latter uses rank correlation test. These tests were performed in STATA and a significance of p < 0.05 in any one of the test indicated the possibility of a bias in the evidence. Appendix 4 gives the results of bias test results. The p-values did not indicate the presence of this form of bias. 3. Results and discussion 3.1. Impact on access: connectivity Fig. 1 provides summary forest plots for different sectors. The overall effect size (1.05) indicates that use of bottom-up approaches has resulted in an increase in connectivity as compared to the topdown approach. However, given the range of the confidence interval, the improvement has not been statistically significant. The sub-group analysis showed that the impact varied by sector and type of facility. However, except in the case of individual toilets, the use of bottom-up approaches did not appear to have a significant impact. The forest plots showed that the effect size was not significantly different from 1, indicating that the beneficiaries were neither better nor worse off after the implementation of bottom-up approaches. In the case of individual toilets, the positive effect of bottom-up approaches could be attributed to (1) the physical or financial contribution made by residents towards the construction of individual toilets, which would be more favourable towards using it and/or (2) the behaviour change efforts by the NGOs who have perhaps been more effective in making residents realise the importance of having the toilets in their homes. NGOs and CBOs model of community engagement often involves community participation in the form of financial contribution or physical effort in the construction of toilets which may ease the financial burden of the residents themselves, thereby facilitating the process of constructing individual toilets. Among all the sectors, the impact had been the least in electricity. A possible reason for the low impact could be due to the centralised nature of electricity supply, which is more suitable for top-down approaches. Government involvement in policy formulation, financing and monitoring of electrification programs is needed in order to play a significant role in improving access (Baruah, 2010; Manzetti and Rufin, 2006; Scott et al., 2005; Shrestha et al., 2008). Since electricity distribution is characterized by limited localisation as compared to water and sanitation services, the involvement of residents could also be limited. The results by type of facility and region are given in Fig. 2. Region wise sub-analysis helps identify whether there are pockets where bottom-up approaches have been more effective (or less beneficial) in the provision of different facilities. Across all four

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

6

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

Fig. 1. Summary results for connectivity.

regions, bottom up approaches appear to have led to a significant improvement in access to individual toilets. However in the case of Africa, the increase was not statistically significant. In water supply (individual connections), the effect size is the highest for African region (1.5) but the results are not statistically significant. A possible reason for this could be the low level of connectivity in the control group; in Africa, the urban government also is largely absent from provision of water and sanitation services. In Central and North America, bottom-up approaches appear to have been ineffective in improving access to individual water connections as well as community toilets. This perhaps could be due to the high cost of connections and ineffective institutions providing service delivery. However, there was only one finding of evidence for Central and North American region, and more studies were needed to understand the reasons for the lack of impact. Fig. 3 shows the results for connectivity by type of slum. Among the 49 findings of evidence for connectivity, 6 were from formal slums, 38 were from informal slums, and 5 were from low-income groups. Bottom-up approaches are apparently the most effective in formal slums. The main reason for this can be attributed to the willingness of NGO's and CBO's to invest in a slum that will not be evicted or demolished. Studies show that improvements in notified slums were implemented much faster than non-notified slums and NGO's were also more active in notified slums (Chandrasekhar, 2005; Kranthi and Rao, 2009). This indicates that de facto or de jure tenure security is an important factor in access to basic services

(Almansi, 2009). In the case of informal slums, the effect size is closer to 1, and the confidence interval range shows no significant difference in connectivity for either of the approaches. The lack of tenure security is often used as a reason by the government to restrict services (government programs and donor assisted programs) to the poor, which results in an absence of investment in poor communities. Moreover, alternate service providers and slum residents may also hesitate to invest large amounts of money in home improvements in an informal slum that is likely to be evicted or demolished (Winayanti and Lang, 2004). As a result, informal slums lack services under both approaches. I-square values indicate that the evidence for formal and informal slums are characterized by considerable heterogeneity. The effect on low-income households has been the least, indicating that when households are not concentrated in a defined location like a slum, the involvement of NGO's and CBO's have not been effective. Fig. 4 gives the summary results on connectivity for community participation. The results show that when there has been some level of community participation, the impact on access appears to be significant. The effect size when there has been community participation was 1.92, and the difference is statistically significant. The inclusive nature of NGOs or a better understanding of the slum realities may be a few of the reasons for why bottom-up approaches appear to deliver better results in slums. Researchers conclude that one of the main reasons for NGO's and CBO's to intervene in the

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

7

Fig. 2. Summary results for connectivity for different regions.

provision of basic services to slums stems from their understanding that slum dwellers needs have been unmet by the government which has resulted in the proliferation of unsanitary practices and health problems and hindered economic activity (Burra et al., 2003; Bapat and Agarwal, 2003; McFarlane, 2009). Additionally, community involvement in the maintenance of infrastructure seems to be better under the bottom-up approach. This is possibly due to the process of management that goes beyond finance to hygiene education, a sense of ownership of community assets and provides a much-needed social space for community members to come together and discuss various issues (WaterAid India, 2008). Alternate service providers also encourage community participation in the planning, design, and maintenance of infrastructure that can yield better results. Fig. 5 shows the impact of participation across different types of slums. While there were no findings of evidence for participation in low-income group, for both formal and informal slums, the results were positive when there was community participation. Particularly, in the case of informal slums, the impact of participation on connectivity is statistically significant. While absence of tenure security may prevent government from providing services, participation from slum residents in the form of self-mobilisation aided

by NGOs and CBOs leads to improved connectivity. The result of participation on formal slums was also positive, with an effect size of 4.06. But the wide confidence interval limits the statistical significance. It can also be said that the tenure security in formal slums enhances participation, which may be in the form of willingness to invest in infrastructure from the slum dwellers and active participation in better design, implementation, and monitoring. 3.2. Impact on access - affordability The indicators of affordability were measured either as the proportion of households who indicated that the services were affordable or the average per unit cost of consumption. When affordability was measure in terms of the proportion of households that indicated that the services were affordable, the pooled effect size (1.11) showed a slight but statistically insignificant increase under bottom-up approaches (Table 2: Panel A(i)). Within the water sector, the evidence was more impressive in the case of community taps as compared to individual, household taps. There are three possible reasons. First, individual connection costs are high and prohibitive for slum dwellers. Therefore, it is economical for a group of households to secure a common connection instead

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

8

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

Fig. 3. Summary results for connectivity by type of slum.

Fig. 4. Summary results for connectivity analysed by level of participation.

of a single family that has to bear the high one-time expenditure (Hardoy and Schusterman, 2000; Weitz and Franceys, 2002; Sohail and Surjadi, 2003). Second, the laying of pipeline networks in the low-income settlements may pose engineering and construction bottlenecks due to narrow roads and congested layouts. Community or public standpipes can be an optimal solution under such

circumstances. Third, NGOs and CBOs prefer to provide community or public facilities as this caters to a large number of residents instead of serving only a few (Bapat and Agarwal, 2003; Burra et al., 2003). Results indicate that bottom-up approaches have been able to leverage the expertise of NGO's and CBO's in making community water supply more affordable. The I-square values also indicate low

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

9

Fig. 5. Summary of evidence for connectivity by slum type and participation.

Table 2 Meta-analysis results for affordability, adequacy, and effort and time. Dimension of Access

Effect Size

Confidence Interval

I-Squared value

P-value

1.11 0.82 1.42 0.39

0.67e1.83 0.33e2.06 1.02e1.97 0.14e1.08

73.1% 81.1% 0.0% 83.4%

0.00 0.668 0.00 0.00

1.11 1.28 1.50 0.82

0.67e1.83 0.74e2.24 0.99e2.26 0.33e2.06

73.1% 0.0% 0.0% 81.1%

0.00 0.00 0.535 0.00

0.92

0.71e1.19

0.0%

0.416

1.06 0.70

0.77e1.47 0.43e1.13

0.0% 0.0%

0.617 0.402

Panel A e Affordability (i) Proportion of households that indicated services were affordable Overall results Water e Individual Tap Water e Community Tap (ii) Mean cost per unit consumed (iii) Type of slum Overall results Formal slum Informal Slum Low-income settlements Panel B: Adequacy (i) Proportion of households that indicated that availability of services was adequate Panel C: Effort and Time (i) Level of effort (ii) Distance travelled

heterogeneity, further enhancing the validity of the results. The meta-analysis results, when evidence was available as the mean cost incurred for unit consumption (Table 2: Panel A (ii)) shows that bottom-up approaches have had a significant impact in making water supply more affordable to slum residents. The effect size (0.39) shows that the mean cost incurred in accessing the services was reduced with the involvement of alternate service providers whose facilities can be accessed at rates that are more affordable than those of top-down regimes. This can be explained by the void in service provisioning in slums which is filled by NGOs and other alternate service providers who may charge a nominal rate when compared to the illegal service providers or private

vendors. When affordability was analysed for different type of slums (Table 2: Panel A (iii)), the effect size shows that bottom-up approaches had the maximum benefit on informal slums in terms of affordability. This could possibly be explained by the unavailability of basic services (poor/low connectivity) in informal slums that forces residents to purchase these services at a much higher cost from private or illegal monopolies. The intervention of NGOs and CBOs results in the provision of supply at a more affordable price (Devas and Korboe, 2000; Hossain, 2012). The I-square values also indicate the absence of heterogeneity, thereby augmenting the validity of the findings.

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

10

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

There has also been an increase in the affordability proportion for formal slums, but this is supported by the presence of single evidence. For low-income households, the mean effect size is lower than 1 (though the difference is not statistically significant), thereby indicating that bottom-up approaches have not led to an increase in affordability. 3.3. Impact on access - adequacy The indicator for adequacy was the proportion of households that indicated that the availability of services was adequate. The results in Table 2, Panel B suggested that in the bottom-up approach, the average proportion of households who found the services adequate was lower in the bottom-up approach. However, the difference was not statistically significant and the effect size was also closer to 1. This indicates that bottom-up approaches did not have any impact on adequacy. 3.4. Impact on access -effort and time The dimension of effort and time is applicable for community water supply and sanitation facilities, where the dwellers have to go to a common point to access the service. Effort and time in literature has been measured either as the distance to be travelled or the time spent to access the facility. The four findings of evidence that we had for this dimension, were grouped into two categories based on the measurement used: (i) As a proportion of households who indicated that the level of effort reduced in bottom-up approach; and (ii) distance travelled to access the services. The mean effect size presented in Table 2, Panel C (i), showed that a higher proportion of households have indicated that they may have needed less effort in accessing the services in bottom-up approach; the difference was however not statistically significant. The involvement of alternate service providers could have a beneficial effect in terms of making access easier for slum dwellers based on the provider's understanding of community needs and the problems faced by them (such as distance to facility) in accessing services (Weitz and Franceys, 2002; Water Aid India, 2008; Burra et al., 2003). The effect size of the evidence on distance travelled by the dwellers (Table 2: Panel C (ii)) shows a reduction in the distance travelled to access when service provision was by alternate service providers. However, given the range of the confidence interval, the null hypothesis of no difference in the distance travelled in bottom up approaches cannot be rejected. The low I-square values indicate that the results are not affected by heterogeneity, possibly because of the fewer findings of evidence for this dimension. 4. Summary and implications This study analyses the effectiveness of bottom-up approaches characterized by decentralisation and involvement of alternate service providers, such as NGOs and CBOs, in the provision of basic infrastructure services for slums and low-income settlements. Findings of evidence from 21 studies that met the inclusion and exclusion criteria were synthesized using meta-analysis. Key results are as follows. Deployment of bottom-up approaches appears to have resulted in an all-round improvement in access. These improvements have only been marginal although there are specific sectors and facilities where the effect appear to be significant. For example, in the case of individual toilets, the adoption of bottom-up approaches has yielded positive results. When analysed by region, the results were significant for Africa as compared to the other regions. The results were also significantly positive for affordability, specifically in the

case of community water connections. While the meta-analysis suggests possible improvements in terms of adequacy and effort and time with bottom-up approaches, the results were not statistically significant. Community participation has emerged as an important factor that can impact the effectiveness of bottom-up approaches. The effect size of community participation was 1.92 and statistically significant. The lack of community participation had an effect size of 0.79. Community participation can help service providers understand the needs of the slum dwellers and subsequently tailor services to suit their requirements. The legal status of the slum determines the type of service provider and the service delivery approach. Formal slums, that are legally recognised and have security of tenure, enjoy a higher degree of connectivity as compared to that of informal slums. Metaanalysis shows that bottom-up approaches are more effective in formal slums (effect size of 2.11) as compared to informal slums (effect size of 0.96). Tenure security provides slum dwellers the safety net to invest in connectivity. Since informal slums are under the threat of eviction or demolition, slum dwellers are reluctant to invest in infrastructure improvements. Taken together tenure security and participation appear to have a multiplier effect on connectivity in slums. The effect size for formal slums where the residents have actively participated in the service delivery process was 4.06, indicating significant improvement in access. However, the effect size for formal slums without community participation was 1.28, indicating that tenure security helps to improve access albeit not at the same level when it is accompanied by community participation. The results of this study have several implications for policy. First, while there has been increasing advocacy on the use of bottom-up approaches to address the needs of the urban poor, blanket adoption of such approaches may not yield the desired results. Bottom-up approaches appear to work better in formal slums and when there is activity participation of the community. Although involving the community may be time consuming, participatory approaches may help improve access by giving the community an opportunity to articulate their needs and contribute to solutions. Therefore, policy makers should seek to include components that will involve the community in the planning and delivery of services. Our results show that bottom-up approaches may be better suited to accomplish community involvement as compared to top-down approaches. Second, providing basic services to citizens in informal and illegal settlements has been a major challenge for governments in LMIC's. Migration from rural to urban areas has led to a proliferation of squatter and illegal tenements in many cities. Governments might restrict such illegal and unplanned tenements, but there is also a moral responsibility of the government to ensure basic living conditions to the citizens. Service delivery by alternative providers in informal slums can be useful because it does not confer the status of tenure security that is usually perceived with government delivery of services. Third, the results obtained from the meta-analysis, including the magnitude of effects, show that bottom-up approaches potentially lead to improvements that go beyond connectivity. NGO's and CBO's proactively take into consideration the immediate needs of the community as well as the maintenance requirements of the project. These factors can lead to positive outcomes with regard to affordability, adequacy, and effort and time, which in turn could lead to better patronage of the facilities and prolonged project life. Finally, our study suggests that the dominant top-down approaches to urban planning may be necessary, but are not sufficient. Enlisting the support of alternate service providers such as NGO's and CBOs and making them a legitimate service partner in

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

last-mile delivery can play an important role in achieving better results. Slum residents are important stakeholders in the process, and a tripartite arrangement that involves the government, the alternate service providers, and the community so that each of these stakeholders contribute in areas of their strength holds tremendous potential in improving access to basic infrastructure services. Acknowledgements Comments and feedback from the anonymous reviewers are gratefully acknowledged. The authors would like to acknowledge the valuable inputs and comments of the Editor which led to marked improvement of the paper. The authors also acknowledge the financial support from DFID under Contract 40077696 for the conduct of this study. Critical inputs on the policy front from Vikram Kapur, former Commissioner, Corporation of Chennai and Raj Cherubal, City Connect, as well as research assistance from Keerthana Sundar is also gratefully acknowledged. Appendix 1. List of inclusion and exclusion criteria

Criteria

11

Gulyani, S.,Talukdar, D.,Mukami, K. R. (2005) Universal (Non) service? water markets, household demand and the poor in urban Kenya. Urban Studies. 42(8): 1247e1274. Hailu, D., Osorio, R. G.,Tsukada, R. (2012) Privatisation and renationalization: What went wrong in Bolivia's water wector? World Development. 40(12): 2564. Hanchett, S., Akhter, S., Khan, M. H.,Mezulianik, S.,Blagbrough, V. (2003) Water, sanitation and hygiene in Bangladeshi slums: An evaluation of the WaterAide Bangladesh urban programme. Environment and Urbanization. 15(2): 43e56. Islam, S. M., Khan, U. N. M. (2013) Access to urban basic services and determinants of satisfaction: A comparison by non-slum and slum dwellers in Dhaka City. Institute of Governance Studies, BRAC University, IGS Working Paper Series No.10/2013, PP.1e30. Israel, D. K. (2007) Impact of increased access and price on household water use in urban Bolivia. The Journal of Environment Development, 16, (58):58e83. Issaka, K. O. (2007) Environmental concerns of poor households in low-income cities: the case of the Tamale Metropolis. Ghana. GeoJournal. 68(4): 343e355. Kayaga, S., Kadimba-Mwanamwambwa, C. (2006) Bridging

Description/Rationale

Language

Only studies published in English were included in this review. However, since English language is one of the preferred languages for researchers, we believe the list of included studies is representative. Year of Publishing Studies that have been published from the year 1999 have been included in the review. This ensured that we synthesized only the recent evidence on this topic. Sectors The coverage of this review encompassed electricity, water, and sanitation sectors. Therefore our search strategy included only those studies that analysed the evidence in these sectors. Outcomes Only studies that analysed the effects of different urban planning approaches on access to basic services were included whereas studies that did not measure the impact on various dimensions of access were excluded. Countries Studies pertaining to developing countries and the ‘transition or emerging’ economies as classified by the World Bank were considered. Broadly, the review covers studies concerning: African countries; countries belonging to parts of South and Central America (Latin America); Asian countries, excluding Japan and the Four Asian Tigers (Hong Kong, Singapore, South Korea and Taiwan); and the transition and emerging economies in Eastern Europe and Central and East Asia. Intervention Those studies that have specifically considered the different attributes of planning in the provision of basic services were included in the review. Type of studies Only original research studies were included in the review. Studies that were reviews of existing studies were not included in the review. However, these reviews formed rich sources of studies for inclusion in this review. Length of Articles that were less than or equal to four pages were not considered for the review, since we felt that such short articles would not provide the Publication necessary information needed for quality appraisal.

Appendix 2. Studies included for meta-analysis Ahmad, S., Choi, M. J., Ko, J. (2013) Quantitative and qualitative demand for slum and non-slum housing in Delhi: Empirical evidences from household data. Habitat International. 38: 90e99. Das, A. K. (2012) End of project evaluation of Madhya Pradesh urban services for the poor programme (MPUSP). End of Project Evaluation Report, Contract Number: 5977, Department for International Development, Government of UK, PP.1e81. Das, M. (2011) End-term impact assessment of the Kolkata urban services for the poor programme. Department for International Development (DFID), India, PP.1e75. Foster, V., Araujo, C. M. (2004) Does infrastructure reform work for the poor? A case study from Guatemala. World Bank Policy Research Working Paper, No. 3185, World Bank, PP.1e76. Galiani, S., Gertler. P., Cooper, R., Martinez, S., Ross, A., Undurraga, R. (2013) Shelter from the storm: Upgrading housing infrastructure in Latin American slums. 3ie, Grantee Final Report, PP.1e49.

Zambia's water service gap: NGO/Community partnerships. Proceedings of the ICE - Water Management, 159 (3):155e160. Kifanyi, G. E., Shayo, B. M. B., Ndambuki, J. M. (2013) Performance of community based organizations in managing sustainable urban water supply and sanitation projects. International Journal of Physical Science, 8, (30):1558e1569. Kranthi, N., Rao, D.K. (2009) Security of tenure and its link to the urban basic services in slums: a case of Hyderabad. The IUP Journal of Infrastructure, VII,(3 & 4):103e113. MdAdbul, B., MdMustak, A., Tofail Md Alamgir, A. (2012) Health status and its implications for the livelihoods of slum dwellers in Dhaka city. Shiree working paper, No. 11, PP.1e40. Mustafa, D., Reeder, P. (2009) People is all that is left to privatize: water supply privatisation, globalization and social justice in Belize City, Belize. International Journal of Urban and Regional Research. 33(3): 789e808. Smith, L., Hanson, S. (2003) Access to water for the urban poor in Cape Town: where equity meets cost recovery. Urban Studies. 40(8): 1517e1548. Stanwix, B. (2009) Urban slums in Gujarat and Rajasthan study of basic infrastructure in seven cities.Mahila Housing SEWA Trust, Ahmedabad, India. PP.1e80.

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

12

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13

Subbaraman, R., Shitole, S., Shitole, T., Sawant, K., Brien, O. J., Bloom, E. D., Patil-Deshmukh, A. (2013) The social ecology of water in a Mumbai slum: Failures in water quality, quantity, and reliability. BMC Public Health, 13,(173): 1e14. Tukahirwa, J. T., Mol. A. P. J.,Oosterveer, P. (2011) Access of urban poor to NGO/CBO-supplied sanitation and solid waste services in Uganda: The role of social proximity. Habitat International. 35(4): 582e591. Zaki, S., Nurul, A. A .T. M. (2009) Does basic services privatisation benefit the urban poor? Some evidence from water supply privatisation in Thailand. Urban Studies. 46(11): 2301e2327.

Appendix 3. Summary of quality appraisal scores of included studies

Strength of evidence

No. of studies

High Medium Low Total

6 14 1 21

Appendix 4. Test results for bias

Evidence

Eggers Test Begg-Mazumdar Test Results (p-value) Results (p-value)

Bottom-up, connectivity Bottom-up, affordability (proportion) Bottom-up, affordability (mean unit consumption costs) Bottom-up, adequacy (proportion) Bottom-up, adequacy (mean consumption) Bottom-up, effort and time (proportion) Bottom-up, effort and time (mean distance travelled)

0.097 0.320

0.109 0.152

0.796

0.806

0.332 e

1.000 1.000

e

1.000

e

1.000

References Almansi, F., 2009. Regularizing land tenure within upgrading programmes in Argentina; the cases of Promeba and Rosario H
abitat. Environ. Urbanization 21 (2), 389e413. Annamalai, T.R., Devkar, G., Mahalingam, A., Benjamin, S., Rajan, S.C., Deep, A., 2016. What Is the Evidence on Top-down and Bottom-up Approaches in Improving Access to Water, Sanitation and Electricity Services in Low-income or Informal Settlements? EPPI-Centre, Social Science Research Unit, UCL Institute of Education, University College London, London. Bakker, K., 2007. Trickle Down? Private sector participation and the pro-poor water supply debate in Jakarta, Indonesia. Geoforum 38 (5), 855e868. Bapat, M., Agarwal, I., 2003. Our needs, our priorities; women and men from the slums in Mumbai and Pune talk about their needs for water and sanitation. Environ. Urbanization 15 (2), 71e86. Baruah, B., 2010. Energy services for the urban poor: NGO participation in slum electrification in India. Environ. Plan. C Gov. Policy 28, 1011e1027. Baud, I., Dhanalakshmi, R., 2006. Governance in urban environmental management:

comparing accountability and performance in multi-stakeholder arrangements in South India. Cities 24 (2), 133e147. Begg, C.B., Mazumdar, M., 1994. Operating characteristics of a rank correlation test for publication bias. Biometrics 50, 1088e1101. Bel, G., Fageda, X., Warner, M.E., 2010. Is private production of public services cheaper than public production? A meta-regression analysis of solid waste and water services. J. Policy Analysis Manag. 29 (3), 553e577. Borenstein, M., Hedges, L.V., Higgins, J.P., Rothstein, H.R., 2011. Introduction to Metaanalysis. John Wiley & Sons. Brooke, P., Smith, W., 2001. Improving Access to Infrastructure Services by the Poor: Institutional and Policy Response. World Bank. Burra, S., Patel, S., Kerr, T., 2003. Community-designed, built and managed toilet blocks in Indian cities. Environ. Urbanization 15 (2), 11e32. Chandrasekhar, S. (2005) Growth of slums, availability of infrastructure and demographic outcomes in slums: Evidence from India. Paper to be presented during the session on urbanization in developing countries at the population association of America, 2005, Annual Meeting, Philadelphia, USA, PP.1e22. €ter, H., 2011. Regulating and deregulating the public Clifton, J., Lanthier, P., Schro utilities 1830e2010. Bus. Hist. 53 (5), 659e672. Connors, G., 2005. When utilities muddle through: pro-poor governance in Bangalore's public water sector. Environ. Urbanization 17 (1), 201e218. Dagdeviren, H., Robertson, A. Simon, 2011. Access to water in slums of sub-saharan Africa. Dev. Policy Rev. 29 (4), 485e505. Dawson, P., 1994. Organisational Change: a Processual Approach. Paul Chapman, London. Devas, N., Korboe, D., 2000. City governance and poverty: the case of Kumasi. Environ. Urbanization 12 (1), 123e136. Devkar, G., Mahalingam, A., Deep, A., Thillairajan, A., 2013. Impact of Private Sector Participation on access and quality in provision of electricity, telecom and water services in developing countries: a systematic review. Util. Policy 27, 65e81. DFID, 2012. Water, Sanitation and Hygiene Portfolio Review. Downloaded from. UKAid Department for International Development. https://www.gov.uk/ government/uploads/system/uploads/attachment_data/file/214187/DFID_ 20WASH_20Portfolio_20Review.pdf. Donna, F.S., Berlin, J.E., Morton, S.C., Olkin, I., Williamson, G.D., Rennie, D., 2000. Meta-analysis of observational studies in epidemiology. JAMA 283 (15), 2009e2012. Egger, M., Davey Smith, G., Altman, D.G., 2000. Systematic Reviews in Health Care: Meta-analysis in Context. British Medical Journal Books, London. Egger, M., Davey Smith, G., Schneider, M., Minder, C., 1997. Bias in meta-analysis detected by a simple, graphical test. Br. Med. J. 315, 629e634. Ghafur, S., 2000. Entitlement to patronage: social construction of household claims on slum Improvement project, Bangladesh. Habitat Int. 24 (3), 261e278. Global Urban Observatory, 2003. Guide to Monitoring Target 11: Improving the Lives of 100 Million Slum Dwellers. UN-HABITAT. Haidich, A.B., 2010. Meta-analysis in medical research. Hippokratia 14 (Suppl. 1), 29e37. Hardoy, A., Schusterman, R., 2000. New models for the privatization of water and sanitation for the urban poor. Environ. Urbanization 12 (2), 63e76. Higgins, J., Thompson, S.G., Deeks, J.J., Altman, D.G., 2003. Measuring inconsistency in meta-analyses. Br. Med. J. 327 (7414), 557e560. Hobson, J., 2000. Sustainable sanitation: experiences in Pune with a municipalNGO-community partnership. Environ. Urbanization 12 (2), 53e62. Hossain, S., 2012. The production of space in the negotiation of water and electricity supply in a bosti of Dhaka. Habitat Int. 36 (1), 68e77. Ibem, Eziyi O., 2009. Community-led infrastructure provision in low-income urban communities in developing countries: a study on Ohafia, Nigeria. Cities 26 (3), 125e132. Kranthi, N., Rao, D.K., 2009. Security of tenure and its link to the urban basic services in slums: a case of Hyderabad. IUP J. Infrastruct. VII (3 & 4), 103e113. Kessides, Ioannis N., 2004. Reforming Infrastructure - Privatization, Regulation; and Competition. World Bank Policy Research Report. World Bank, Washington, DC. Kingston, R., (1998) Web based GIS for public Participation decision Making in the UK, Paper presented at the Empowerment, Marginalization, and Public Participation GIS Meeting, Santa Barbara, California.
n, M., McGranahan, G., 2006. Informal Water Vendors and the Urban Poor. Kjelle International Institute for Environment and Development, London. Manzetti, l., Rufin, C., 2006. Private Utility Supply in a Hostile Environment: the Experience of Water, Sanitation and Electricity Distribution Utilities in Northern Colombia, the Dominican Republic and Ecuador. Sustainable Development Department Best Practices Series. Inter-American Development Bank, Washington, D.C, pp. 1e63. McFarlane, C., 2009. Sanitation in Mumbai's informal settlements: state, slum and infrastructure. Environ. Plan. A 40 (1), 88e107. MDG, 2013. Millennium Development Goals Report. United Nations, New York. Mundial, B., 2004. World Development Report 2004: Making Services Work for Poor People. Banque mondiale, Washington, DC. Nelson, L., Dowling, P.J., 1998. Electricity industry reform: a case analysis in Australia. J. Organ. Change Manag. 11 (6), 481e495. Nijman, J., 2008. Against the odds: slum rehabilitation in neoliberal Mumbai. Cities 25, 73e85. OECD, 2013. Financing Water and Sanitation in Developing Countries: the Contribution of External Aid. Downloaded from. http://www.oecd.org/dac/stats/ Brochure_water_2013.pdf. Otiso, K., 2003. State, voluntary and private sector partnerships for slum upgrading

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002

S. Narayanan et al. / Utilities Policy xxx (2017) 1e13 and basic service delivery in Nairobi City, Kenya. Cities 20 (4), 221e229. Pettigrew, A.M., 1987. Context and action in the transformation of the firm. J. Manag. Stud. 24 (6), 649e670. Roma, E., Jeffrey, P., 2011. Using a diagnostic tool to evaluate the longevity of urban community sanitation systems: a case study from Indonesia. Environ. Dev. Sustain. 13 (4), 807e820. Rosenthal, R., 1991. Meta-analytic Procedures for Social Research, third ed. Sage, Newbury Park, CA. Rothstein, H.R., Sutton, A.J., Borenstein, M., 2006. Publication Bias in Meta-analysis: Prevention, Assessment and Adjustments. John Wiley & Sons. Scott, N., McKemey, K., Batchelor, S., 2005. Energy in Low-income Urban Communities, Barriers to Access to Modern Energy in Slums. Final technical report, Contract Number R8146, pp. 1e161. Shrestha, R.M., Kumar, S., Martin, S., Dhakal, A., 2008. Modern energy use by the urban poor in Thailand: a study of slum households in two cities. Energy Sustain. Dev. XII (4), 5e13. Sohail, M., Surjadi, C., 2003. Pubic Private Partnership and the Poor, Drinking Water Concessions: a Study for Better Understanding Public-private Partnership and Water Provision in Low Income Settlements. WEDC, Loughborough University, UK, pp. 1e48. Sterne, J.A.C., Gavaghan, D., Egger, M., 2000. Publication and related bias in metaanalysis: power of statistical tests and prevalence in the literature. J. Clin. Epidemiol. 53, 1119e1129.

13

UNCHS, 2006. Enabling Shelter Strategies: Review of Experience from Decades of Implementation. United Nations Centre for Human Settlements, Nairobi. UN-Habitat, 2012. State of the World's Cities 2010/2011: Prosperity of Cities. Rutledge. UN-Habitat. (2002) Expert Group Meeting on Urban indicators held in Nairobi, November 2002. Report available at: http://www.unhabitat.org/programmes/ guo/documents/EGM_finalreport4Dec02.pdf). UN-Habitat, 2003. Water and Sanitation in the World's Cities - Local Action for Global Goals. Earthscan, London. United Nations Conference on Trade and Development. World Investment Report 2008, 2008. Transnational Corporations, and the Infrastructure Challenge. UN. USAID, 2004. Innovative Approaches to Slum Electrification, Bureau for Economic Growth, 20523. Agriculture and Trade U.S Agency for International Development, Washington, D.C, pp. 1e137. WaterAid India, 2008. Tiruchirappalli Shows the Way: Community-municipal Corporation-NGO Partnership for City-wide Pro-poor Slums Infrastructure Improvement, pp. 1e38. Weitz, A., Franceys, R., 2002. Beyond boundaries extending services to the urban poor. Asian Dev. Bank. 1e108. Winayanti, L., Lang, H.C., 2004. Provision of urban services in an informal settlement: a case study of Kampung Penas Tanggul, Jakarta. Habitat Int. 28 (1), 41e65.

Please cite this article in press as: Narayanan, S., et al., Delivering basic infrastructure services to the urban poor: a meta-analysis of the effectiveness of bottom-up approaches, Utilities Policy (2017), http://dx.doi.org/10.1016/j.jup.2017.01.002