(Mis)Conceptualising Empowerment: Flaws in Community-driven Development Project Design

Abstract

Development institutions have adopted models of community development that prioritise empowerment. However, many of these models have transformed a collective, process-oriented understanding of empowerment into one that focuses on individual outcomes and results all too often in project failure. This article presents a case study of a World Bank community-driven development (CDD) project implemented in agrarian settlements in north-eastern Brazil. Using mixed methods, the article finds that because this CDD approach failed to conceptualise power differences, stemming from education, income and the rural–urban divide, it overlooked the disadvantage marginalised communities would have when interacting with technical agencies and soliciting contractors. The power differences created by these inequities enabled those very technical agencies and contractors intended to support CDD subprojects to take advantage of participant communities through providing low-quality services and products. The project also failed to provide quality technical assistance for a duration sufficient to develop the skills community members required to carry out successful, productive subprojects. These findings suggest that CDD projects need to incorporate a broader conception of empowerment that addresses group-level power differences to transform the relations between participants and those on whom they depend, and thereby contribute to project success.

Keywords

Participatory development community-driven development empowerment Latin America capacity power

Introduction

Over the 1980s, 1990s and 2000s, top-down, state-driven and neoliberal approaches towards development were criticised for producing lacklustre growth, for failing both to reach the poorest and to value local knowledge and for lacking pluralism and sustainability (Chambers 1983; Chang 2003; Hickey and Mohan 2005; Narayan-Parker and Patel 2000). Responding to this criticism, development institutions began to emphasise human development as a measure of economic development (Mohan and Stokke 2000; Pieterse 1998). The focus on human development coincided with the use of a more decentralised development approach that aimed to increase community involvement (Fukuda-Parr et al. 2002; Gillis et al. 2001; Veltmeyer 1997).

One such approach was the World Bank’s model of community-driven development (CDD). Empowerment, which it defines as ‘expanding the assets and capabilities of poor people to participate in, negotiate with, and hold accountable institutions that affect their lives’, is a principle of this approach (Binswanger-Mkhize et al. 2010, 4). CDD resulted from development institutions modifying an existing, and more radical, participatory development approach to work within their theoretical and practical structures—and moulded the concept of empowerment to do the same (Cornwall 2008). The form of empowerment that emerged tended to focus on individual outcomes, while minimising inter-group power differences and the conflict that could emerge from the process of marginalised groups becoming empowered (Mohan and Stokke 2000). It also minimised the transformative political potential of participatory development (Williams 2004). The question then arises, what are the repercussions of this more limited theoretical conception of empowerment in the context of a CDD project?

In order to address that question, this article presents a case study of a World Bank CDD project, São José Agrário (SJA), implemented in the state of Ceará, north-eastern Brazil. SJA was a spin-off of the larger, state-wide São José II project (a World Bank Rural Poverty Alleviation Project). SJA involved setting up productive and infrastructure subprojects in agrarian settlements to increase household income, improve well-being and strengthen social capital (World Bank 2006). Ceará has been experimenting with CDD-type projects since the 1980s (Coirolo and Lammert 2009), leading to significant experience with such projects. Additionally, this SJA project provides a valuable example of a CDD project intimately connected to social movements and located in agrarian settlements with high levels of social capital. Households in these communities were fairly homogeneous, had similar goals and had significant experience with collective action and organising. In sum, these were ideal communities for a CDD project. As such, the case studies provide clarity in identifying project design weaknesses, because community collective action problems were minimised.

The CDD project design was expected to cultivate empowerment through providing communities the power to choose subprojects from a menu, control and account for subproject funds and contract with and oversee technical agencies (Coirolo and Lammert 2009, 105). This empowerment is manifested as decision-making power, voice and the ability overcome state and market failures (Mansuri and Rao 2012). The goal of SJA was to increase income and assets, and access to water and electricity (World Bank 2009). The CDD funding parameters for SJA required that the participants create complex subproject proposals and meet accounting requirements (World Bank 2001b). However, according to the author’s survey and interviews, participants had low levels of education and significant illiteracy levels. Consequently, CDD requirements presented a challenge to the participants. Moreover, they were expected to engage in new types of productive activities destined for the market; in essence, they were expected to become entrepreneurs. However, participant surveys revealed that they had little business experience from which to draw. Prior to moving to the settlements, most had been either subsistence farmers or landless workers (author’s interviews and survey).

In the case studies presented here, agrarian settlements needed significantly more skill and resources than they had to design and carry out their selected subprojects. To resolve this problem, the CDD project design relied on technical agencies, while also assuming that participants would monitor and enforce agreements with contracted companies to provide quality materials and subprojects. However, this article proposes that because the CDD project did not conceptualise group power differences, it failed to predict the difficulty participants would have in enforcing contracted companies and technical agencies to meet community priorities. Agrarian settlers, as peasants, occupied a position of subordination within their economic, social and political structures (Akram-Lodhi and Kay 2009), which hindered their ability to interact effectively with government officials, technicians and contractors. I find that this problem was further magnified when the communities relied on private technical agencies, whose foremost goal was profit. In some cases, the technical agencies used their power to take advantage of the communities and provided low-quality subprojects and materials.

The participants needed these technical agencies because they had to learn new skills and encountered significant obstacles, all the while having little of their own experience, social connections or education they could draw on. Consequently, the participants required mentoring throughout the maturation of the subproject. Nonetheless, in the case studies presented here, communities received technical assistance at most only through the first year, even though productive subprojects took multiple years to mature. Together, these flaws in project design often led to subproject failure. A technician summarised the situation in the larger São José II project:

In the end, the great majority of these [sub]projects failed. These are [sub]projects of little reach, ... some of the [sub]projects are practically idle, the infrastructure has been built . . . and nothing is working. This was very bad for all of us that work with sustainable rural development.¹

In the subsequent iteration of this project, the World Bank has required that the communities contribute a greater percentage of the subproject cost (discussed in more detail in the section ‘Response and Conclusion’) (World Bank 2012). This renewed approach suggests that the World Bank attributed at least a portion of the subproject failure to problems of moral hazard—participants took on too much risk and did not have enough buy-in (personal investment in the project expressed as cash or in-kind contributions)—rather than to flaws in project design.

This article argues that the CDD project design’s concept of empowerment focused on individuals and outcomes and ignored group power differences and processes and, consequently, reproduced dependencies rather than creating empowerment. Participant dependence on technical assistance, and the ability of technical agencies to leverage that dependence, in some cases, led to poorly constructed subprojects. This article contends that project design, combined with an inadequate period of training and support, contributed to subproject failure and led to the World Bank’s misunderstanding of the reasons for that failure.

Participatory development projects must conceptualise a process of empowerment that addresses group power differences and focuses on the process of building critical consciousness and collective power to achieve empowerment and successful outcomes. Therefore, the theoretical concept of empowerment employed in this CDD subproject—individually located and measured through outcomes—constrained participant capacity growth, critical consciousness and organising power to confront disadvantageous power dynamics and solve subproject problems. This study contributes to the literature identifying the limitations of an individual, outcome-focused conception of empowerment. It also highlights how the complexity of the bureaucratic processes limits participants’ ability to be successful and provides potential lessons to practitioners. Practitioners should take into account the differences in power stemming from income, education and the rural–urban divide which limit the ability of communities to assume an equal role with other development partners, such as contractors and private technical agencies.

This article is organised as follows. The next section provides a brief literature review and discussion of the concept of empowerment used in development projects and notes connections to key features of my case study. The subsequent section introduces the case study background and methods. The fourth section reviews how the project design generated dependence rather than empowerment. In the final section, the article analyses the World Bank’s response to subproject failures and concludes.

Conceptualising Empowerment

Empowerment lies at the heart of participatory development, which depends on the agency of participants in the development process (Vainio-Mattila 2000). The concept of empowerment has been elaborated on in various contexts, including psychology, feminist studies and development studies (Cornwall 2016; Speer et al. 2001; Zimmerman 1995). The popularity of the concept has resulted in numerous uses and definitions, leading to ambiguity in meaning (Batliwala 2007; Ibrahim and Alkire 2007; Orbach 2011).

Radical feminist social scientists and development studies scholars have provided useful analyses and theorisations of the concept of empowerment in development discourse. Women’s empowerment has emphasised the importance of critical consciousness and relational (including intra-household) power hierarchies within the broader context of the society, including those stemming from the intersections of class, race, caste, et cetera (Batliwala 2007; Cornwall 2016; Kabeer 1999; Leder et al. 2017; Rowlands 1997).

The concept of empowerment meaningful to radical participatory development prioritises process and group. Radical participatory development grew out of the alternative development approach of the 1970s–1990s; it was influenced by postmodern thought and advanced critiques of modernist development approaches (Hickey and Mohan 2005). In this analysis, empowerment is a collective process that demands at least some emphasis on political and social change achieved through struggle and conflict (Laverack 2001).

Both feminist and radical participatory development approaches centre power in their analyses. They often draw on the Freirian concept of conscientizecão, sometimes called ‘critical consciousness’, to confront ‘structures of oppression’ in which the oppressed transform the institutions of society, including those of the state and the market (Cleaver 1999; Freire 1970; Mohan and Stokke 2000; Ward 2010). Rowlands (1995, 1997), theorising empowerment in the context of gender and development initiatives in Honduras, has provided a schema of power: power within, power to, power with and power over. Recognising and confronting internalised power structures, power from within, is similar to critical consciousness and is the basis for transformational change (Rowlands 1997). Power to is akin to Kabeer’s (1999) concepts of resources and agency and Sen’s (1985) theory of capabilities. In the context of the larger family, community and societal relationships, these represent the ability to set goals, make decisions and act on them to achieve desired outcomes without domination (Cornwall 2016; Kabeer 1999). Power with is the power to take collective action with others, and power over is the power to dominate, where some people have control or influence over others (Rowlands 1995).

I classify definitions of empowerment along two lines: first, based on whether empowerment is considered a process or an outcome (Laverack 2001) and, second, based on whether empowerment is conceptualised at the individual level (such as decision-making within the household) or at a group level (such as occurring through collective action). In the radical feminist tradition, empowerment is considered a process that often prioritises, first, the individual and relational position of women and, second, the group. This emphasis comes out of the unique position in which women operate—they are members of families, intensely interdependent and often responsible for the care of others. Radical participatory development focuses on the process and often prioritises the group or collective concerns first, but it recognises the importance of critical consciousness at the individual level for building collective action.

When development institutions took up the torch of empowerment as a prong of their participatory development approach, they were criticised for co-opting the meaning of participatory development and empowerment and turning an intensely political development strategy focused on power into an apolitical set of tools with measurable and codifiable outcomes to prove their efficacy (Chhotray 2004; Cooke and Kothari 2001; Hickey and Mohan 2005; Kapoor 2002; Mosse 2001). The need for development institutions to measure (count) the impact of their interventions moves their focus from process to outcomes, leading to a limited conception of empowerment emphasising economic effects (Batliwala 2007; Carter et al. 2014; Cleaver 1999; Dill 2009). Interestingly, several World Bank reports theorise process and societal context in empowerment (Alsop et al. 2005; Eyben et al. 2008; Mansuri and Rao 2012). Mansuri and Rao (2012, 88) discuss the importance of the collective in working together to achieve outcomes and redress state and market failures. This limited conception of collective power does not address the very important aspects of inter-group power arising from income, rurality or education. These characteristics are precisely what would make it difficult for the participants to hold the state accountable or to solve market failures. The CDD project reviewed in this article did not evaluate the projects beyond the increase in incomes, access to electricity and water and, for productive subprojects, their economic rate of return (World Bank 2009).

In worst-case scenarios, empowerment is used to appeal to a greater morality; rather than striving towards transformation, development institutions often prioritise efficiency and reproducible technique (Cleaver 1999; Toomey 2011). Empowerment thus becomes a technocratic exercise with an itemised quality; subprojects are chosen from menus, and funds are disseminated in tranches that require extensive documentation and reporting (Chhotray 2004). As such, development institutions have taken the ‘power out of empowerment’ (Batliwala 2007).² They have failed to explicitly challenge existing societal power hierarchies (Moore 2001) and so have provided an incomplete concept of empowerment. By emphasising the individual economic empowerment outcomes at the local level, generally via increases in assets or market participation, development institutions have de-emphasised the importance of critical consciousness and collective action, as well as the role the state plays in maintaining power hierarchies (Dodman and Mitlin 2013; Mohan and Stokke 2000; O’Hara and Clement 2018).

When development institutions ignore society-level power hierarchies—Rowlands’ power over—they are unable to conceptualise and create systems and institutions that would shift these hierarchies (Cornwall 2016; Rowlands 1995). Instead, they can disempower the poor by reproducing dependency and constructing institutions that are not inclusive or genuinely participatory (Dill 2009; Orbach 2011). For example, in my projects, the World Bank and the state government required a level of education and technical expertise the community members did not have, making them reliant on a host of others. In addition, because there were no group-level theorisations of power, the large power differences between the rural, less educated and poorer community members and the generally urban, more educated and richer others, with whom the community members were required by project structures to interact, were not confronted.

Background and Method

I chose the World Bank CDD project SJA, located in Ceará, Brazil, because this project came about through a social movement’s—the Landless Workers’ Movement (MST)—occupation of state offices and demand for project funds. The MST won these funds and the right to choose the settlements that would participate in the project. Therefore, the project was conducted under the advantageous conditions of high levels of community organisation—agrarian settlements affiliated with social movements and agricultural workers’ unions. Organised communities with a history of collective action seemingly have the greatest chance of making CDD projects work, which allows for easier identification of project shortcomings that cannot be attributed to a community’s lack of organisation and coordination.

These communities overcame some important problems often present in CDD projects, namely clientelism and free riding. Even so, half of my case study subprojects had failed or been put on hold. The World Bank and the Department of Agrarian Development (Secretaria de Desenvolvimento Agrário, SDA), responsible for overseeing the São José projects, classified subprojects as infrastructure or productive (World Bank 2009). Six of the eight subprojects were productive subprojects, which were expected at implementation to continue into the foreseeable future. These included beekeeping (apiculture), irrigation subprojects for fruit trees and vegetable crops, growing of capim (a grass feed for livestock) and a cashew plantation (see Table 1). Of these six productive subprojects, three were either non-existent or on hold when I visited.

My mixed-methods study is based on 14 months of fieldwork conducted during July–August 2011 and July 2012–June 2013. As noted earlier, the SJA project was a spin-off programme of the World Bank Rural Poverty Alleviation Project, locally named São José II. SJA provided funds for 180 subprojects to be implemented in agrarian settlements affiliated with the MST, 163 of which were approved and implemented. The maximum value of a project was USD 50,000. A total of USD 15 million was available to fund the SJA project, which covered the subprojects, training and overhead administration.

From the MST, I obtained a list of the SJA subprojects by subproject type, subproject cost, settlement and the number of participants. I matched this with official lists of agrarian settlements in the state by the number of households, date of officially becoming a settlement and municipality. In selecting from the 163 subprojects to study, I prioritised minimum variation due to geographic location (including micro-climate and political factors), the year in which the land occupation became a settlement and the number of families within settlements. Applying these criteria, I selected eight subprojects, seven of which were distributed across two municipalities, Canindé and Quixeramobim. A similar microclimate—semi-arid (the sertão)—prevails in both municipalities, which are also geographically proximate and contain many subprojects. In addition, the seven selected subprojects were based in settlements established between 1998 and 2002. This meant they were constructed under similar land reform policies and were of similar ages—at least 10 years old by the beginning of the study—affording them sufficient time to experiment with and establish a range of institutions. As for the number of families, I selected 10–30 as the range to ensure settlements of similar sizes. The seven settlements included all settlements meeting the above criterion in the municipalities of Canindé and Quixeramobim. Finally, I included an eighth settlement located in a different municipality (Itapipoca), named Settlement 1 in Table 1. This settlement was included because of its strong affiliation with the MST enhanced my analysis of the relationship between the MST and the other settlements.

Table 1.

São José Agrário Subprojects.

Settlement	Municipality	Subproject	Type	Year	Cost	Existing
1	Itapipoca	Cashew	Productive	2008	R$70,000	Yes
2	Quixeramobim	Fence	Infrastructure	2008	R$46,000	Yes
3	Caninde´	Irrigation	Productive	2009	R$9,700	No
4	Quixeramobim	Capim/shed/fence	Productive/infrastructure	2010	R$50,000	Yes
5	Caninde´	Apiculture	Productive	2008	R$58,800	On hold
6	Caninde´	Reservoir	Infrastructure	2009	R$90,000	No
7	Caninde´	Irrigation	Productive	2010	R$82,000	No
8	Quixeramobim	Tractor	Productive	2010	R$80,000	Yes

Source: The author.

Note: ‘Municipality’ is the location of the settlement. ‘Subproject’ describes the subproject received by the community. Subprojects can be classified as either productive or infrastructure subprojects. ‘Year’ is the year the subproject was implemented. ‘Settlement’ is the name I gave the settlement. ‘Cost’ describes the total grant given for the subproject. ‘Existing’ describes if the subproject was still in existence at the time of my visit. ‘On hold’ describes a subproject that was not currently functional when I visited but which some of the settlers expected to start again sometime in the future. Settlement 4’s subprojects were a combination of three smaller subprojects—including a capim (a feed for cattle), a storage shed and a fence. These composed both productive and infrastructure subprojects.

I conducted interviews with a variety of actors associated with the settlements and with the SJA subprojects. Interviews conducted in the settlements with subproject participants were structured. All subproject-participant interviewees were asked the same questions, but follow-up probes could differ. I used structured interviews with settlement members, because they shared many of the same conditions and roles: living in a land reform settlement and being a community member. Through using structured interviews, I ensured I was consistent in how and what I asked community members, to enable a more accurate comparison of their answers. In the first two settlements, I interviewed 12 and 14 households, respectively, close to half the households in each community. According to the literature on the method of interview saturation, between 12 and 30 people is the point at which redundancy is reached (depending on the context, use of expert interviews and emphasis on pragmatic or total redundancy) (Guest et al. 2006; Trotter et al. 2001). In my context, I reached data saturation, the point at which no new theme or significant variation in experiences and answers occurred (Glaser and Strauss 1967; Trotter et al. 2001), after interviewing around six households in Settlements 1 and 2, but I continued up through around half the households, to 14 and 12, respectively (see Table 2). This relatively small sample size to redundancy occurred because the interviewees came from similar backgrounds, engaged in similar work, lived together in a similar context and were jointly involved in the same subproject. Interviewees were purposively sampled for variation in age, gender and leadership roles within each settlement (Ritchie and Lewis 2003). These interviews covered the SJA subprojects and their participation in these, as well as their crops and animals, other projects they may have received, access to health and education, income and participation in community work.³ I conducted 39 interviews with settlement members; the interviews lasted between 45 minutes and 1 hour (see Table 2).

Table 2.

Survey and Structured and Semi-structured Interviews.

Survey	No.	Structured Interviews	No.	Semi-structured Interviews	No.
Settlement 1	0	Settlement 1	14	Landless Workers’ Movement Representatives	4
Settlement 2	0	Settlement 2	12	Agricultural Workers’ Union Representatives	7
Settlement 3	12	Settlement 3	2	Sa˜o Jose´ Agra´rio Technicians	4
Settlement 4	10	Settlement 4	3	Technical Agencies Representatives	2
Settlement 5	19	Settlement 5	2	Land Reform Agency Representatives	5
Settlement 6	10	Settlement 6	2	Religious Organization Representatives	1
Settlement 7	25	Settlement 7	2
Settlement 8	17	Settlement 8	2
Total	93	Total	39	Total	23

Source: The author.

Note: Structured interviews lasted between 45 minutes and 1 hour. Semi-structured interviews averaged 1–2 hours.

Interviews with technicians, government officials, MST representatives, agricultural union representatives and other actors were semi-structured. I used a semi-structured interviewing technique to flexibly adapt question themes to the interviewees’ roles and their relationship to the subproject communities, which varied greatly. I adopted a strategy of purposive and snowball sampling. I had a list of organisations and actors who interacted with and supported the subproject communities. I contacted the organisations and actors for interviews. At the end of the interviews, I inquired into the possibility of conducting additional interviews with others they considered appropriate. I conducted 23 interviews via this method; the interviews lasted 1–2 hours (see Table 2). These interviews covered the subprojects, questions regarding the viability of the settlements, agricultural techniques, land reform and access to health and education.

Using this information, I created a survey instrument that I applied to all households in the six remaining settlements in Canindé and Quixeramobim, for a total of 93 households (see Table 2). The survey lasted between 45 minutes and an hour and a half. It covered individual factors for each family member, such as age, gender, race, literacy, education and relationship to the head of the household. It then included questions about household income, assets, employment, subproject participation and collective settlement work. The survey had a response rate of greater than 90%. In these settlements, I also conducted interviews with at least two current or former settlement leaders, following the same questions as the interviews in the first two settlements. When possible, I checked my empirical data against the national, state and rural components of the Brazilian household survey of 2012 (Instituto Brasileiro de Geografia e Estatistica [IBGE] 2012).

I spent around a week in each settlement, attending meetings and social events, and sometimes helping with household work. This provided me additional insights into the complexities of the settlements and the subprojects (Basole and Ramnarian 2016). In particular, variations in income became more apparent, as did relationships and fissures within the communities. I also observed general assembly meetings and meetings on other projects currently being implemented. Additionally, I was able to observe collective settlement work. My observations enhanced the interview and survey results through adding information not easily obtained through those methods.

All interviews were recorded, transcribed and coded. I used a combination of descriptive coding and hypothesis coding. While I went into my study with some background knowledge and several hypotheses, I sought to hear what the participants said beyond what I expected. For this reason, I began my first round of coding with descriptive coding for topics (Saldana 2012). I then broke relevant topics down into sub-topics and ran queries for problems identified by participants (Bazeley and Jackson 2013). Additionally, I used hypothesis coding, wherein I tested whether the answers I expected were correct (Saldana 2012). In particular, I used this method to explore problems with elite capture, free riding and literacy.

Through including all settlements with subprojects that met my criteria, in the two municipalities and conducting a census survey in six of these settlements, I was able to obtain comprehensive descriptive statistics at the individual settlement level. These data are representative of the settlements that met my criteria in these two municipalities. They are not representative of settlements outside the criteria upon which I made my selections, such as larger settlements or settlements in other regions, which may have had different experiences.

Empowerment or Dependence

Accessing Subprojects

Although the São José II project (the larger, state-wide umbrella project) and the SJA project targeted the poor (World Bank 2009), I found they demanded significant education and literacy, constituting an obstacle to participation. SJA participants often did not have the necessary literacy, education and skills to create the subproject proposal or plan, implement and manage the SJA subprojects independently.

Not surprisingly, my sample had low literacy rates and educational levels. Before joining the settlements, most individuals had been landless workers or subsistence farmers (moradores), who comprise the poorest groups in the region. My census survey revealed a 29.5% illiteracy rate (295 people, 15 years and older) compared with 7.5% for Latin America and the Caribbean as a whole, 9% for Brazil and 18% for Ceará (all for the year 2012) (IBGE 2012; UNESCO Institute for Statistics 2012). This finding is similar to two other surveys done of rural populations in Ceará about 10 years earlier, one finding 32% and the other 33% of settlement heads of households to be illiterate (Filho et al. 2001; Leite et al. 2004). The average duration of education of my sample of all participants aged 18 years and older was 5 years, with a median of 4 years of education (see Table 3).

Table 3.

Years of Education by Age Group.

Age (years)	Median	Mean	Standard Deviation	No. of People
18–27	10.5	9.01	2.81	76
28–37	4	4.78	3.33	46
38–47	4	3.9	2.89	58
48–57	1	2.28	2.73	32
58–67	1	1.72	2.79	25
68–77	0	1.53	2.29	17
78–87	0	0	0	5
All ages	4	4.9	4.02	259

Source: The author.

Note: I grouped people by age groups of 10 years and then provide the median, mean and standard deviation of education in each group.

In São José II, communities applying for a CDD subproject were required to create a subproject proposal that outlined the subproject design, including the needed inputs, a budget and a plan for operations and maintenance (World Bank 2009). Illustrative of the required literacy is the following excerpt from the subproject proposal submission guidelines for the São José III Word Bank project (the most recent iteration of the project begun in 2013), which breaks down proposal expectations:

Furthermore, proposals should include information on: (a) market demand for product commercialization; (b) availability of and demand for inputs required for production; (c) viability study; (d) organizational and administrative capacity of the proposing organization; (e) logistics and strategy for commercializing the products; (f) technical design (description of necessary works, technical specifications, budget and list of suppliers of the required equipment); (g) operational framework and sustainability strategy; (h) environmental aspects and specific measures to prevent or minimize environmental impacts; (i) management plan; (j) financial and accounting management; and (k) water availability and source (for irrigation investments). (World Bank 2012, 27)

To some extent, this problem was alleviated in SJA. As discussed previously, the MST organised an occupation by agrarian settlers of the Secretary of Agrarian Development of Ceará and demanded access to São José II project funds, which they won and which became SJA. According to interviews, the MST chose which settlements would receive funds based on their participation in the occupation.⁴ Thus, settlements were able to access the subprojects without first submitting a proposal. While the participants were still required to follow all the larger São José II project requirements, including the eventual submission of the subproject proposal, they were not required to submit the proposal prior to being chosen for a subproject and received some support in proposal creation.

Accounting Standards and Financial Management

As part of the management of the subproject, participants were responsible for recording how subproject funds were used. Due to the low levels of education and literacy, the required power to meet the subprojects’ accounting standards was explicitly reported to be a challenge in interviews with members of the municipal-run agricultural workers union, the MST, by participants in Settlements 3 and 5, and by technicians.⁵ The Department of Agrarian Development technicians in charge of the São José projects consistently reported that the communities faced problems working with banks, sourcing equipment and supplies, administering money, keeping financial records and presenting the required documents and signatures to the government officials in charge of the projects.⁶

The problem was aggravated by a difference in accounting standards understood by the communities and the Department of Agrarian Development regarding the organisation and completeness of the accounts. The associations received the subproject money in a bank account, dedicated solely to the subproject, in three tranches. To get the second and third tranches, they were required to report and show the receipts that accounted for the preceding tranches of funding.⁷

According to a technician from the Department of Agrarian Development, the community associations in charge of the subprojects often lost receipts or forgot to account for them, resulting in incomplete financial records.⁸ Some settlers mentioned keeping of accounts to be a problem. For example, the president of Settlement 3 said, ‘It [keeping accounts] is a big challenge. Because of the large degree of illiteracy. Until a person understands what keeping accounts is, understands numbers, there is a lot of difficulty’.⁹ The president of Settlement 7 responded that this was not a problem for them and that keeping accounts was simple: it only required noting down expenses and keeping track of receipts.¹⁰ When I relayed this comment to a Department of Agrarian Development technician, he replied that overall, for most communities, the level of record-keeping was less thorough than what the state offices required:

Yesterday they [the subproject participants] came to submit a report, already it is the third time these people have come to submit a report for access to funds. It is for cashews. And they forget some simple signatures. We remind them, return the documents, and they come back again, but it takes them a long time to return. Then they bring one document [when in fact] we asked for a whole list of documents. They bring one, but forget the others. It seems that grassroots groups still do not value documentation, they still think that keeping accounts is something that is done by experts. [They think] that they cannot take on this process. In a way, I think to a certain extent the institutions [technical agencies] that work with the farmers failed to empower and train them for this type of activity.¹¹

While the government needed this paperwork trail to ensure that funds were being directed appropriately, many subproject participants, as well as the MST, perceived this list of requirements as impediments to successfully implementing the subprojects on time. Hold-ups in tranche releases of subproject funds—either because the settlers were not presenting the correct documentation or because the state government was not meeting its timeline—meant settlers were sometimes beginning new subprojects out of season. For example, in Settlement 1, there was a delay in receiving the resources to contract for the cashew trees. When they did receive the resources to plant the trees, it was during the dry season instead of the rainy season in a subproject with no irrigation.

The emphasis on accounting for funds and paperwork is a result of the outcome-oriented nature of the SJA project that emphasised the need to measure inputs and outputs. These requirements presented little intrinsic importance or value to the communities. They were complied with to the best of the communities’ abilities because the World Bank had power over them to set the frame for subproject participation.

Technical Agencies

Fostering Dependence on Technical Agencies

The SJA project design relied on technical agencies to bridge the gap between the skills community members needed to create and carry out subprojects and the skills they had (World Bank 2009). In all the eight subprojects I studied, the communities received extensive technical assistance in creating their subproject proposals. This skill gap was not unique to my case studies. In a World Bank case study of four productive subprojects of the São José II project, the authors found that the business plan¹² in all four subprojects depended on the help of outside agencies: either technical agencies, the state technical unit or—for two of the cases—the Bank of Brazil (52).¹³ Dill (2009, 728), in a study on community-based participation in Tanzania, similarly found that international non-governmental organizations stepped in to assist the communities in the registration process and the creation of the required constitution for the community-based organisations useful for communities hoping to access participatory development funds.

Participants, of course, desired that the subproject be well-designed. The literature on CDD bases much of the rationale for this type of development project on the incentive that participants have to monitor and enforce the effective use of subproject funds by contracted agencies, eliminating corruption (Wong 2012). I found that because some of my SJA case study subprojects were prepared by private technical agencies, a principal–agent problem was created. The reason was that the goals of the communities and those of the private technical agencies were at odds. The communities wanted the most cost-effective, highest-quality subproject; the technical agencies wanted the largest profit margin for the least effort. This is precisely because the income of the technical agency is a cost to the community (Brett 2003). As Brett (2003) explains, it is essential in these situations that the community have real leverage over the technical agency to enforce compliance.

However, the case study subproject proposals were technical documents, often incomprehensible to the layperson, resulting in asymmetric information that forced the community to rely on the technical agency for creation and interpretation. Technicians arrived in the communities from the privileged position of generally having come from higher incomes, more urban areas and greater education, giving them power over the community members. Additionally, being the technician gave them expert power. The communities had little leverage to monitor and enforce their priorities on technicians and contractors.

A recurrent theme emerging from interviews was that technical agencies took advantage of the communities through spending less time and resources elaborating and accompanying the subproject to widen their profit margin. This behaviour sometimes resulted in low-quality subprojects based on inferior materials, and technical agencies faced no repercussions for subproject failure once subprojects were approved. One such example was Settlement 7’s failed irrigation subproject, which encountered problems from its beginning. The 18-foot-deep well the technical agency dug filled with high-salinity water. Instead of continuing to drill to reach freshwater, drilling a new well or revising the subproject altogether, the technical agency considered the well satisfactory and continued with the subproject. The high-salinity water caused two problems. First, the irrigation tubing diameter was too small, such that salt and mineral build-up inside the tubes blocked water flow. Second, the crops chosen—vegetables and banana and papaya trees—did not respond well to the high-salinity water and died. Furthermore, the technical agency had also chosen a pump too weak to push the water through the tubing to irrigate the whole area.

According to one settlement member, when a state technician (not a technician from the company that had designed the subproject) visited the settlement, he said the irrigation subproject had been executed incorrectly, particularly in that the tubing going to the plant root was not in place. The former president of Settlement 7 concluded,

The issue I found is that so many [sub]projects that come, the companies that win the right to give technical assistance, they are only interested as long as they are receiving their payment and afterward they do not help … not even in the beginning. The company won the right to give assistance, both to create a proposal and to outline the [sub]project correctly. This business … came here very few times. It did not follow up.¹⁴

This same reflection concerning private technical agencies was relayed in interviews by others in the municipality of Canindé, primarily by members of Settlement 3 and a member of the public technical agency for the municipality, Ceará Technical Assistance and Rural Extension Business (Empresa de Asistencia Técnica e Extensão Rural do Ceará).¹⁵ They reported that technicians from private agencies would make only superficial visits—just enough to meet the requirements of the subproject.

Rather than empowering participants, the project design obstructed participation and created community dependence on technical assistance. The focus on project outcomes, rather than on the process, presented a series of issues. One, the communities were not capacitated to create these subproject proposals (the power to was not cultivated), so their creation was outsourced to technical agencies. Two, the subproject proposals were technical documents that were difficult to understand and mostly unread by the communities. For example, in all but one settlement, Settlement 5, the community leaders were unable to locate either the written subproject proposal or the subproject guidelines. The World Bank had power over the communities, requiring them to develop documents to meet World Bank systems, for which the communities seemingly had little use or need. Lastly, the communities were unable to hold accountable the private technical agencies with whom they contracted, because the latter also had power over the communities stemming from group power differences of education and income.

Inadequate Duration of Technical Assistance

Despite the problem of private technical agencies taking advantage of the communities, technical assistance was essential to subproject success, particularly for productive subprojects. Surveys showed that communities were inexperienced in the productive subprojects (they did not have the power to), having had little exposure to the new types of crops or to the transportation and marketing of goods. Most participants had very little practical management knowledge. Of the 93 households I surveyed, only one household head had managed workers, and only one other had experience with commercial activities relevant for managing a subproject. The lack of experience meant that a primary component of empowerment was to assist the subproject participants in learning how to carry out the subprojects and overcome subproject problems. In the case studies presented in this article, the duration of technical assistance was too brief to teach the skills participants needed to carry out the subprojects successfully. As such, the subproject focus on design and standardisation hindered participant improvements in capacity (Mishra 2016).

In part, this disconnect occurred because productive subprojects were following an infrastructure time frame for technical accompaniment. This characteristic appears to be a general problem noted in previous CDD projects that placed greater emphasis on infrastructure subprojects (Mansuri and Rao 2012, 109). The São José subprojects’ movement from infrastructure towards productive subprojects can be seen in the increasing total number of productive subprojects and the total funds dedicated to them. Of the São José I subprojects, launched in 1995, only 1% of the total subprojects were productive subprojects; in São José II, launched in 2001, 18% were productive subprojects, and São José III, launched in 2013, estimates that 76% of all subprojects would be productive subprojects (see Table 4; World Bank 2001a, 2009, 2012).

In the settlements I studied, the productive subprojects received technical assistance only during the first year. This was sufficient for infrastructure subprojects but inadequate for productive subprojects, which took multiple years to mature. To better illustrate this point, I have classified productive and infrastructure subprojects into four stages in Table 5. As the table illustrates, productive subprojects were still evolving over stages 3 and 4. I categorised the Incubation stage as the time in which the productive subproject required participant labour, sometimes participant monetary contributions, but did not yet have marketable output. In the Output stage, there was output that had to be transported and marketed, again requiring participant labour and monetary contributions. Only in the Output stage would the productive subproject contribute directly to income increases for the participants. Thus, for infrastructure subprojects, the time between implementation and outcome was relatively short, whereas, for productive subprojects, the time between implementation and outcome was much longer.

Table 4.

World Bank Funding for Community Subprojects by Type.

	No. of Total Projects	Total Cost (US$ million)	Infrastructure Subprojects		Productive Subprojects		Social Subprojects
	No. of Total Projects	Total Cost (US$ million)	No. (% total no.)	Cost in US$ Million (% total cost)	No. (% total no.)	Cost in US$ Million (% total cost)	No. (% total no.)	Cost in US$ Million (% total cost)
SJI	3,057	92.9	2,354 (77)	71.5 (77)	30 (1)	0.9 (1)	673 (22)	20.5 (22)
SJII	2,932	104.5	2,391 (81.55)	82.3 (79)	531 (18.11)	19.4 (19)	10 (0.34)	2.8 (3%)
SJIII	585	120.0	140 (24)	50 (42)	445 (76)	70 (58)	0 (0)	0 (0)

Source: World Bank (2001a, 2009, 2012).

Table 5.

Infrastructure and Productive Subproject Stages.

Stages	1	2	3	4
Infrastructure	Approval	Implementation	O&M	None
Productive	Approval	Implementation	Incubation	Output

Source: Author’s typology and data.

Note: I categorised the incubation stage as the time during which the productive subproject required participant labour, sometimes participant monetary contributions, but did not yet have marketable output. In the output stage, there was output that had to be transported and marketed, again requiring participant labour and monetary contributions. Only in the output stage would the productive subproject contribute directly to increases in participants’ income. Thus, for infrastructure subprojects, the difference in time between implementation and outcome was relatively short, whereas for productive subprojects, the time gap could be significant.

O&M = Operation and maintenance.

While technical assistance for the infrastructure subprojects was provided through their most critical stage—implementation—productive subprojects lacked technical assistance through their equally critical Incubation and Output stages (see Table 5). This was a particular problem for the productive subprojects, in which many new skills had to be learned during the Incubation and Output stages, such as developing, growing, marketing and transporting the product.

Not only was the time frame short for technical assistance, but also the subprojects only had a small portion of their funds dedicated to technical assistance. According to one state-level technician, this was between 2% and 3%.¹⁶ The technical agencies explained to the subproject participants that their percentage cut from the subproject was not enough to cover technical assistance beyond one or two subproject visits. Thus, the private technical agencies that elaborated these subprojects provided little actual technical accompaniment for the subproject. The following two case studies outline the problems the subprojects faced in the Incubation and Output stages, at which time they were not receiving technical assistance.

As an illustration, Settlement 3 implemented an irrigation subproject for fruit trees, primarily bananas, in 2009 for R$9,700. The money covered paying the technical agencies that provided assistance, buying the irrigation pump and setting up the irrigation system. Settlement 3 planned to sell the bananas at the nearest market in another town. The subproject went well for about a year. Over that period, they planted and watered the banana trees. The banana trees started producing bananas, but the settlers were unable to transport the bananas to the market. They needed a truck but only had motorcycles. The settlers were unable to overcome these obstacles and instead sold their bananas to a middleman. At around the 1-year point, the participants grew disillusioned. Settlement 3’s subproject failed because the revenue they obtained for the bananas (partly because they were selling through a middleman) was insufficient to cover the cost of electricity needed to run the irrigation pump. When I arrived, the irrigation pump was in disrepair, they had ceased to water the banana plants, and the settlement had no plans to continue in the future. Only approximately 10% of the banana trees had survived. The settlers were unable to solve the problem of transportation to the market because, one, they did not have the power to find a solution and, two, they were in an unequal power relation with the middleman who, due to greater income and assets, had power over the settlers in defining the price of the bananas.

Settlement 5 faced a similar problem with their beekeeping subproject: the settlers had a difficult time marketing the honey. Additionally, they did not like honey and so did not consume their production. According to my survey and interviews, they had been unable to sell any honey following the first year of honey production.¹⁷ Over the following years, the region faced a drought. Lacking flowers, the bees had to be fed. Again, the settlers were not making any revenue to cover the costs of production. The subproject was on hold when I visited in 2013. Settlement 5 had initially wanted a different project. When the majority voted for the beekeeping subproject, eight households left the subproject. The beekeeping subproject was a result of a technician who heavily favoured apiculture subprojects and, via his power over the community members, convinced them to try this subproject.¹⁸ This eroded the strength of the community members to organise together (power with). Additionally, they were not capacitated to solve the myriad of problems that appeared (power to).

In both of these cases, technical assistance could have supported the settlers’ endeavours to find solutions to these problems, supporting their power to. For example, in Settlement 3, technical assistance could have helped the settlers find an alternative way to get their bananas to the market and cut out the middleman; technical assistance could also have facilitated their applying for an available grant that provided lower electric rates outside of peak times. In Settlement 5, technical assistance could have helped the settlers locate suitable markets to sell their honey.

Response and Conclusion

The CDD approach examined in this study understood empowerment to occur through subproject participation. First, individual empowerment was expected to result through participant decision-making (choice of subproject), control over subproject funds and contracting technical agencies (Coirolo and Lammert 2009). Second, carrying out the subproject was expected to develop new skills and capacities, resulting in increased income, assets or access to water or electricity (World Bank 2009). CDD assumes that participation in subprojects would increase individual and community capacity while improving subprojects because participants have the greatest incentives to foster high-quality projects (Mansuri and Rao 2012). Unfortunately, CDD project designs and capacity development programmes generally ignore the importance of societal power structures—often deriving from group-level characteristics (Bockstael 2017). The sources of power I have highlighted in this article stem from education, income and the rural–urban divide. These power asymmetries can be seen as manifestations of deep-rooted inequalities. In ignoring group power relations, the CDD approach reproduces rather than transforms these existing power relations (McCARTHY et al. 2017). These case studies, therefore, provide an example of a problem noted to be a common issue of the CDD approach.

The subproject outcomes were prioritised over the process, seen in the emphasis on completing the subprojects, rather than on building participant capacity to carry out the subprojects without requiring significant dependence on others. As such, the power within (to see themselves as equals, building critical consciousness of societal power relations) that underlies the power to (capacity to do) and the power with (effective collective action) was not supported (see Rowlands 1995, 1997). This resulted in many problems: the participants were dependent on experts (all settlements), they may have taken on a subproject that was not their top priority (Settlement 5), they had difficulties in accounting for funds (all settlements), they found it difficult to oversee technical agencies (Settlement 3, Settlement 6 and Settlement 7), and they did not have the capacity to maintain or solve subproject problems once the technical agencies left (Settlement 3, Settlement 5 and Settlement 7).

The communities chose their subproject. However, this power to was constrained, first, by their choosing only from a menu of possible subprojects and, second, by the technicians’ power over the participants based on education and income. The fact that the project design emphasised project proposals, reporting and accountability mechanisms beyond the capacities of the participants exemplifies the constraints. These requirements stemmed from the development institutions’ need to measure outcomes rather than focusing on building critical consciousnesses to confront power relations with the technical agencies or creating skills. The problem of dependence was exacerbated when the participants relied on private technical agencies. CDD project designs expect the community to oversee the technical agencies to ensure subproject proposals are properly designed and implemented. These case studies illustrate the difficulty that the poor, rural and relatively less educated subproject participants had in enforcing their priorities on private technical agencies and contractors, and in monitoring the agencies’ actions precisely because of these deep-seated differences in group power.

I argue that CDD blindness to group power differences created subprojects that put the participants in a dependent position, allowing private technical agencies to take advantage of them (wield power over). The dependence on technical agencies also contributed to another problem. Participants in productive subprojects generally were engaged in a new production method or crop type. Were they to be successful in achieving production, they then had to find an appropriate market for the good, transport it to a market and address any equipment failures, pests or other problems that arose. Participant inexperience meant that the duration of the technical assistance for productive subprojects, around 1 year, fell significantly short of enabling communities to learn by doing (power to). Thus, when the technical assistance ended, often after the subproject implementation, the settlers were unable to solve the aforementioned problems. Instead of empowering the individuals involved, the project design created participant dependence on others for project design, project implementation and project problem solving.

Of the eight subprojects that I visited, four no longer existed. Of the six productive subprojects that were originally supposed to generate income directly or indirectly, only the capim subproject had produced indirect income via the limited production of feed for settlement-owned livestock. In any case, subproject outcomes ranged from outright failure to fairly dismal, with two exceptions: the capim and the cashew subprojects. The cashew plantation had yet to mature and still had the potential to be income-generating.

The World Bank and the state development officials did not foresee such problems because the SJA project did not conceptualise group-level power differences. As such, it did not recognise the power technical agencies and contractors had over the settlers due to their subordinated position within society. Additionally, the SJA project prioritised outcomes over the process and so did not foment the critical consciousness (power within) that could have supported collective action (power with) to confront these power hierarchies.

In contrast to my findings, the World Bank and the Department of Agrarian Development have at least partly attributed the subproject failures in São José II and SJA to moral hazard, as they have increased the ‘buy-in’ for communities to participate in the subprojects. The moral-hazard argument assumes participants knowingly took on subprojects that were too risky. This assessment relies on the assumption that participants can accurately forecast subproject costs and benefits. However, in the case of productive subprojects, it is quite difficult for participants to predict costs and benefits, because these subprojects consist of both new types and new methods of production.

In 2013, a new iteration of the São José II project began, the Ceará Rural Sustainable Development and Competitiveness Project—locally named São José III. São José III requires greater buy-in by project participants. São José II and SJA required that the participants provide 10% of the subproject grant, which could be in cash or in kind (labour). São José III requires that the participants provide 20% of the subproject grant, of which 10% must be in cash (including bank loans), and 10% can be in kind or in cash (World Bank 2012).

Requiring additional ‘buy-in’ may exclude the poorest from participating in the subprojects, as they would be unable to provide 10% of the subproject cost in cash, whereas in many cases, they can provide it in labour (in kind). In fact, in the context studied, participants lost significant previous buy-in from failed subprojects in terms of loss of the time, labour and money they put into the subprojects, as well as the costs from the death of plants and animals and from equipment breakage. Greater buy-in would not increase the ability of the participants to predict subproject costs and benefits accurately. However, it would increase the risk to participants regarding what they would lose if the subproject fails.

This study contributes to the practice of participatory development through providing the following two reflections. One, conceptualising empowerment to include elements of group power relationships is not an empty exercise but has real implications for project design and success. Such a concept recognises the potential dependence created by project structures (such as assuming that participants can effectively discipline those with more income or education) and addresses this problem to transform the relations between participants and those on whom they depend, thereby contributing to subproject success. For example, if subprojects did not depend on complex subproject proposals, and instead communities that had already implemented similar subprojects assisted new communities, the power relations between the two would be more similar, and plans and information would be more transparent. Second, subproject emphasis on meeting bureaucratic reporting requirements not only was an obstacle to meaningful participation by participants but also created another hurdle through forcing participants to rely on external help to create documents they did not use beyond meeting subproject conditions.

While this study cannot be generalised beyond the communities interviewed and surveyed, problems identified in this project may also occur in other projects because of the formulaic nature of CDD project designs. Further research is needed to conclude if the creation of dependence rather than empowerment and the reliance on an inadequate timeline of technical assistance for productive subprojects is a recurring feature of the CDD process, as well as being characteristic of all the São José´ projects.

Footnotes

Acknowledgments

I am indebted to João Paulo de Souza, J. Mohan Rao, Mwangi wa Githinji, Krista Harper, Gerald Epstein, Francisco Amaro Gomes de Alencar, Ki Kim and Joo Yeun Suh for their helpful comments on previous versions of this article. I am responsible for all errors.

Declaration of Conflicting Interests

Funding

The author received no financial support for the research, authorship and/or publication of this article.

Notes

ORCID iD

Jessica Carrick-Hagenbarth

References

Akram-Lodhi

A. Haroon

Kay

Cristóbal

. 2009. ‘Peasant Livelihoods and the Agrarian Question’. In Peasants and Globalization: Political Economy, Rural Transformation and the Agrarian Question , edited by Akram-Lodhi

A. Haroon

Cristo´bal

Kay

, 3–34. New York: Routledge.

Alsop

Ruth

Bertelsen

Mette

Holland

Jeremy

. 2005. Empowerment in practice: From analysis to implementation . The World Bank.

Basole

Amit

Ramnarian

Smita

. 2016. ‘Qualitative and Ethnographic Methods in Economics’. In Handbook of Research Methods and Applications in Heterdox Economics, edited by Frederic S. Lee and Bruce Cronin, 135–64. Northampton, Massachusetts: Edward Elgar Publishing.

Batliwala

Srilatha

. 2007. ‘Taking the Power Out of Empowerment—An Experiential Account’. Development in practice 17(4–5):557–65.

Bazeley

Patricia

Jackson

Kristi

eds. 2013. Qualitative Data Analysis with NVivo . 2nd edition. Los Angeles: SAGE Publications.

Binswanger-Mkhize

Hans P.

Jacomina

P. de Regt

Spector

Stephen

. 2010. Local and Community Driven Development: Moving to Scale in Theory and Practice: New Fronters of Social Policy: 5330 . Technical Report. Washington, DC: World Bank.

Bockstael

Erika

. 2017. ‘Critical Capacity Development: An Action Research Approach in Coastal Brazil’. World Development 94:336–45.

Brett

Eduard A.

2003. ‘Participation and accountability in development management’. The Journal of Development Studies 40(2):1–29.

Carter

Jane

Sarah

Byrne

Schrader

Kai

. 2014. ‘Learning About Women’s Empowerment in the Context of Development Projects: Do the Figures Tell Us Enough?’ Gender & Development 22(2):327–49.

10.

Chambers

Robert

. 1983. Rural Development: Putting the Last First . Hoboken, NJ: John Wiley.

11.

Chang

Ha-Joon

. 2003. ‘Introduction’. In Rethinking Development Economics, edited by Ha-Joon Chang, 1:1–18. New York: Anthem Press.

12.

Chhotray

Vasudha

. 2004. ‘The Negation of Politics in Participatory Development Projects, Kurnool, Andhra Pradesh’. Development and Change 35(2):327–52.

13.

Cleaver

Frances

. 1999. ‘Paradoxes of Participation: Questioning Participatory Approaches to Development’. Journal of International Development 11(4):597–612.

14.

Coirolo

Luis

Lammert

Jill

. 2009. Rural Poverty Reduction in Northeast Brazil: Achieving Results through Community Driven Development . Technical Report. Washington, DC: World Bank.

15.

Cooke

Bill

Kothari

Uma

. 2001. Participation: The New Tyranny? London: Zed Books.

16.

Cornwall

Andrea.

2008. ‘Unpacking “Participation”: Models, Meanings and Practices’. Community Development Journal 43(3):269–83.

17.

Cornwall

Andrea

. 2016. ‘Women’s empowerment: What works?’ Journal of International Development 28(3):342–59.

18.

Dill

Brian

. 2009. ‘The Paradoxes of Community-Based Participation in Dar es Salaam’. Development and Change 40(4):717–43.

19.

Dodman

David

Mitlin

Diana

. 2013. ‘Challenges for Community-Based Adaptation: Discovering the Potential for Transformation’. Journal of International Development 25(5):640–59.

20.

Eyben

Rosalind

Naila

Kabeer

Cornwall

Andrea

. 2008. Conceptualising Empowerment the Implications for Pro Poor Growth: A Paper for the DAC Poverty Network Brighton: Institute of Development Studies.

21.

Filho

Hildo Meirelles de Souza

Buainain

Antonio Marcio

Mello

Carolina Junqueira Homem de

Silveira

José Maria da

Magalhaes

Marcelo M.

. 2001. ‘Does Community-Based Self Selection of Land Reform Beneficiaries Work? An Assessment of the Brazilian Cedula da Terra Pilot Program’. Economia 2(1):3–45.

22.

Freire

Paulo

. 1970. Pedagogy of the Oppressed . 30th anniversary edition. New York: Continuum.

23.

Fukuda-Parr

Sakiko

Lopes

Carlos

Malik

Khalid

, eds. 2002. Capacity for Development: New Solutions to Old Problems . New York: Earthscan/UNDP.

24.

Gillis

Malcolm

Shoup

Carl

Sicat

Gerardo P.

. 2001. World Development Report 2000/2001—Attacking Poverty . Washington, DC: World Bank.

25.

Glaser

Barney

Strauss

Anselm

. 1967. ‘The discovery of grounded theory. 1967’. Weidenfield & Nicolson, London: 1–19.

26.

Guest

Greg

Bunce

Arwen

Johnson

Laura

. 2006. ‘How Many Interviews Are Enough?’ Field Methods 18(1):59–82.

27.

Hickey

Sam

Mohan

Giles

. 2005. ‘Relocating Participation within a Radical Politics of Development’. Development and Change 36(March):237–62.

28.

Ibrahim

Solava

Alkire

Sabina

. 2007. ‘Agency and Empowerment: A Proposal for Internationally Comparable Indicators’. Oxford Development Studies 35(4):379–403.

29.

Instituto Brasileiro de Geografia e Estatistica (IBGE). 2012. Pesquisa Nacional por Amostra de Domicilios (PNAD) . Rio de Janeiro, Brazil: Instituto Brasileiro de Geografia e Estatistica.

30.

Kabeer

Naila

. 1999. ‘Resources, Agency, Achievements: Reflections on the Measurement of Women’s Empowerment’. Development and Change 30(3):435–64.

31.

Kapoor

Ilan

. 2002. ‘The Devil’s in the Theory: A Critical Assessment of Robert Chambers’ Work on Participatory Development’. Third World Quarterly 23(1):101–17.

32.

Laverack

Glenn

. 2001. ‘An Identification and Interpretation of the Organizational Aspects of Community Empowerment’. Community Development Journal 36(2):134–45.

33.

Leder

Stephanie

Clement

Floriane

Emma

Karki

. 2017. ‘Reframing Women’s Empowerment in Water Security Programmes in Western Nepal’. Gender & Development 25(2):235–51.

34.

Leite

Sergio

Heredia

Beatriz

Medeiros

Leonilde

Palmeira

Moacir

Rosangela

Cintraso

. 2004. Impactos dos assentamentos: um estudo sobre o meio rural brasileiro [Settlement Impacts: A Study of Rural Brazil] . Brasilia: Ministario do Desenvolvimento Agrário, INCRA.

35.

Mansuri

Ghazala

Rao

Vijayendra

. 2012. Localizing Development: Does Participation Work? Washington, DC: World Bank Publications.

36.

McCARTHY

J. F.

Steenbergen

D. J.

Warren

. 2017. ‘Community Driven Development and Structural Disadvantage: Interrogating the Social Turn in Development Programming in Indonesia’. The Journal of Development Studies 53(12):1988–2004.

37.

Mishra

Pradeep Kumar

. 2016. ‘Managing International Development Projects: Case Studies of Implementation of Large-Scale Projects in India’. International Journal of Rural Management 12(1):4–26.

38.

Mohan

Giles

Stokke

Kristian

. 2000. ‘Participatory Development Empowerment: The Dangers of Localism’. Third World Quarterly 21(April):247–68.

39.

Moore

Mick

. 2001. ‘Empowerment at last?’ Journal of International Development 13(3):321–9.

40.

Mosse

David

. 2001. ‘“People’s Knowledge” Participation and Patronage: Operations Representations in Rural Development’. In Participation: The New Tyranny? edited by Cook

Kothari

, 16–35. London: Zed Books.

41.

Narayan-Parker

Deepa

Patel

Raj

. 2000. Voices of the Poor: Can Anyone Hear Us? Vol. 1. Washington, DC: World Bank Publications.

42.

O’Hara

Corey

Clement

Floriane

. 2018. ‘Power As Agency: A Critical Reflection on the Measurement of Women’s Empowerment in the Development Sector’. World Development 106:111–23.

43.

Orbach

Ariella

. 2011. ‘Rethinking Participation Power: Rethinking Reflections on Local Autonomy’. Canadian Journal of Development Studies 32(2):196–209.

44.

Pieterse

Jan Nederveen

. 1998. ‘My Paradigm or Yours? Alternative Development, Post-development, Reflexive Development’. Development and Change 29(2):343–73.

45.

Ritchie

Jane

Lewis

Jane

. 2003. Qualitative Research Practice: A Guide for Social Science Students and Researchers . SAGE Publications.

46.

Rowlands

. 1995. ‘Empowerment Examined’. Development in Practice 5(2):101–107.

47.

Rowlands

. 1997. Questioning Empowerment: Working with Women in Honduras . Dublin: Oxfam.

48.

Saldana

Johnny

. 2012. The Coding Manual for Qualitative Researchers . 2nd. Los Angeles: SAGE Publications.

49.

Sen

Amartya

. 1985. Capabilities and Commodities . Amsterdam and New York: North-Holland.

50.

Speer

Paul W.

Jackson

Courtney B.

Peterson

Andrew

. 2001. ‘The Relationship Between Social Cohesion N. and Empowerment: Support New Implications for Theory’. Health Education & Behavior 28(6):716–32.

51.

Toomey

Anne H.

2011. ‘Empowerment and Disempowerment in Community Development Practice: Eight Roles Practitioners Play’. Community Development Journal 46(2):181–95.

52.

Trotter

Robert T.

Needle

Richard H.

Goosby

Eric

Bates

Christopher

Merrill

Singer

. 2001. ‘A Methodological Model for Rapid Assessment, Response, and Evaluation: The RARE Program in Public Health’. Field Methods 13(2):137–59.

53.

UNESCO Institute for Statistics. 2012. Data: Literacy rate, by age group (% of people ages 15 and above): Latin America and the Caribbean . http://uis.unesco.org/

54.

Vainio-Mattila

Arja

. 2000. ‘The Seduction and Significance of Participation for Development Interventions’. Canadian Journal of Development Studies 21(1):431–46.

55.

Veltmeyer

Henry

. 1997. ‘Decentralisation as the Institutional Basis for Community-Based Participatory Development: The Latin American Experience’. Canadian Journal of Development Studies 18(2):303–25.

56.

Ward

Paul

. 2010. ‘Participatory Development in Jamaica: Does It Work in Practice?’ Social and Economic Studies 59(4):167–96.

57.

Williams

Glyn

. 2004. ‘Evaluating Participatory Development: Tyranny, Power and (Re)Politicisation’. Third World Quarterly 25(3):557–78.

58.

Wong

Susan

. 2012. What Have Been the Impacts of World Bank Community-Driven Development Programs? CDD Impact Evaluation Review and Operational and Research Implications . Washington, DC: Citeseer.

59.

World Bank. 2001a. Implementation Completion Report on a Loan in the Amount of USD 70 Million to the State of Ceará for a Rural Poverty Alleviation Project: Report No: 23397 . Technical Report. World Bank, Sustainable Development Department, Brazil Country Management Unit, Latin America and Caribbean Region, Brasília, Brazil.

60.

World Bank. 2001b. Project Appraisal Document on a Proposed Loan in the Amount of USD 37.5 Million to the State of Ceará for a Rural Poverty Reduction Project (Brazil Rural Poverty Reduction Program: Report No: 21434-BR . Technical Report. World Bank, Brazil Country Management Unit, Environmentally, Socially Sustainable Development Sector Management Unit, Latin America, and Caribbean Region, Brasília, Brazil.

61.

World Bank. 2006. Project Paper on a Proposed Additional Financing (Loan) in the Amount of USD 37.5 Million to the State of Ceará with the Guarantee of the Federative Republic of Brazil for the Ceará Rural Poverty Reduction Project: Report No: 36062-BR . Technical Report. Brazil Country Management Unit, Environmentally, Socially Sustainable Development Sector Management Unit, Latin America, and Caribbean Region, Brasília, Brazil.

62.

World Bank. 2009. Implementation Completion and Results Report on a Loan in the Amount of USD 37.5 Million Equivalent and an Additional Financing Loan in the Amount of USD 37.5 Million Equivalent to the State of Ceará for the Rural Poverty Reduction Project: Report No. ICR00001216 . Technical Report. World Bank, Sustainable Development Department, Brazil Country Management Unit, Latin America and Caribbean Region, Brasília, Brazil.

63.

World Bank. 2012. Project Appraisal Document on a Proposed Loan in the Amount of USD 100 Million to the State of Ceará with a Guarantee of the Federal Republic of Brazil for the Ceará Rural Sustainable Development and Competitiveness Project (Projeto de Desenvolvimento Rural Sustentavel—Projeto São José III): Report No: 66407-BR . Technical Report. World Bank, Sustainable Development Department, Brazil Country Management Unit, Latin America and the Caribbean Region, Brasília, Brazil.

64.

Zimmerman

Marc A.

1995. ‘Psychological Empowerment: Issues and Illustrations’. American Journal of Community Psychology (New York) 23(5):581–99.