Motivations for slum dweller social movement participation in urban Africa: a study of mobilization in Kurasini,Dar es Salaam

Abstract

This paper examines what motivates the participation of African slum⁽¹⁾ dwellers in urban social movement activities. This issue is analyzed through a case study of grassroots mobilization around evictions in Kurasini ward, Dar es Salaam, Tanzania. The paper uses an analytic narrative approach to account for patterns in participatory behaviour, drawing on both quantitative and qualitative data gathered through interviews with 81 slum dwellers. The study shows that, contrary to the expectations of movement leaders, property owners were significantly more likely than renters to participate in a risky and time-consuming mobilization effort. The study identifies three factors that favoured owner participation: the nature of expected payoffs from participation; greater belief in their efficacy of action; and greater connection to place.

I. Introduction

This paper examines the motivations for African slum dweller participation in urban social movement activities through a study of mobilization undertaken by the Tanzania Federation of the Urban Poor (TFUP), the national slum dweller movement of Tanzania. As a starting point, the paper develops and tests three hypotheses concerning motivations for participation in risky and time-consuming social movement activities. Specifically, the paper examines whether payoffs, identity and connection to place influence decisions to participate. The first two factors are part of a long-standing debate on participation in civic and organizational life,⁽²⁾ while the last has been relatively unstudied in the context of social movements.⁽³⁾

There is increasing focus in the international development and planning literature on participatory and bottom-up approaches to development,⁽⁴⁾ including a growing awareness of the role that social movements play in mobilizing the poor.⁽⁵⁾ While efforts have been made to understand the conditions under which movements as a whole arise and succeed,⁽⁶⁾ there has been less research on the micro-dynamics of individual participation in these movements, particularly in the developing world. Recognizing this, this study examines the dynamics of social movement participation in an increasingly important locus for grassroots organizing, namely urban slums.⁽⁷⁾ Given the relatively small sample size (n = 81), which was defined by the number of people participating in the activities in question, an analytical narrative approach, combining formal and narrative methods and drawing upon both quantitative and qualitative data, was adopted.⁽⁸⁾ The paper’s hypotheses formalize a set of a priori expectations based on past research and intuition developed in the field, and provide a framework for identifying themes that are then investigated in greater detail through a more nuanced analysis of qualitative data gathered through interviews.

While centred on a single case study and notwithstanding the fact that African cities demonstrate great variation in local politics, culture and physical form, the findings are likely to be relevant in a wide variety of similar contexts. This is because the wider economic and political forces shaping Dar es Salaam and around which TFUP mobilizes – including increasing demand for the marginal land on which the poor live – also influence other cities in Africa and elsewhere in the developing world.⁽⁹⁾ Enhancing the applicability of this study’s findings to other cities, the form of grassroots mobilization investigated here – a grassroots enumeration based on a household survey – is increasingly employed by slum dweller groups around the world.⁽¹⁰⁾ Since more than 70 per cent of urban residents in many sub-Saharan African countries live in slums,⁽¹¹⁾ the study’s findings are widely relevant, particularly in contributing insights into the associational life of the urban poor.

II. The Case Study: Kurasini, Dar Es Salaam

The paper centres on a study of TFUP mobilization around evictions in Dar es Salaam’s Kurasini ward, home to more than 34,500 people. Plans for the redevelopment of Kurasini have circulated since 2001, when the city council and Ministry of Lands, Housing and Human Settlements Development published the Kurasini Area Redevelopment Plan.⁽¹²⁾ This document argued for the redevelopment of unplanned settlements surrounding the city’s port and increasing fuel storage capacity in the area. The first wave of evictions in Kurasini took place in October 2007.

TFUP is associated with Slum Dwellers International (SDI), a global network of slum dweller federations, and works with the Centre for Community Initiatives (CCI), a Tanzanian partner NGO that provides capacity-building support to the national movement.⁽¹³⁾ Established in 2004, TFUP was, at the time of the Kurasini eviction, the most recent national affiliate to join the SDI network. TFUP is built on a nationwide network of 30 community savings groups, with a total of 1,700 members.⁽¹⁴⁾ Prior to the eviction, Kurasini was home to 300 of these members, distributed over seven savings groups.⁽¹⁵⁾

TFUP is ostensibly organized around savings activities and the granting of small loans to members. Members participate in savings groups, which promote income generation and facilitate loans for housing and infrastructure improvements.⁽¹⁶⁾ The savings groups also provide a means of connecting members and sharing information. In addition, TFUP has strong movement dimensions, with the savings groups serving as the basis for more political aspects of slum dweller mobilization around local land, water and sanitation issues. In Kurasini, some TFUP members actively participated in efforts to minimize the impacts of evictions in the ward, while others continued to be involved in less political aspects of the movement, such as attending weekly savings group meetings. The principal mobilization effort undertaken by TFUP members around the eviction in Kurasini was a grassroots enumeration of neighbourhoods that were to be evicted.

Enumerations consist of a population census and comprehensive mapping of plots and households.⁽¹⁷⁾ Enumerations have several goals: they demonstrate that marginalized communities have the capacity for self-organization; they generate data that give a tangible identity to slum residents; and they serve as the basis for lobbying. The enumeration in Kurasini collected demographic data on residents, including age, household size and income, and also recorded the dimensions of plots and individual structures. The data were recorded in a database and each plot and structure was added to a map developed by the enumeration team.

Accepting that the Kurasini eviction would go ahead, TFUP hoped to lobby government for a grant of land for community resettlement.⁽¹⁸⁾ TFUP movement organizers attempted to convince authorities in municipal and national agencies responsible for land administration that they should provide a grant of land to which evictees could move before they were forced to leave their homes in Kurasini. Within Tanzanian central government, the primary regulatory agency with respect to urban land management and allocation is the Ministry of Lands, Housing and Human Settlements Development. Because this ministry can allocate lands for specific purposes, it was the primary target of TFUP’s lobbying efforts to secure a grant of land for resettlement.

Enumerations have been employed widely by SDI affiliates around the world and are a well-established element of the SDI methodological toolkit.⁽¹⁹⁾ TFUP organizers learned about the methodology and its application through exchanges with SDI affiliates in India, South Africa and Kenya.⁽²⁰⁾ The primary way in which TFUP members could participate in the enumeration process was by serving as “enumerators”, the name given to individuals responsible for collecting demographic and housing data from community residents. The Kurasini enumeration organized in 2007 was the second enumeration undertaken by TFUP.

Since the enumeration entailed considerable risk and was time consuming, the decision to participate serves as a powerful indicator of “meaningful”, rather than tacit, movement participation. The enumeration was also physically challenging, involving long working days outdoors. Due to political and social tensions over land and housing in Tanzania, members who served as enumerators often faced hostile reactions from residents who suspected they were working with government, developers or land speculators.⁽²¹⁾

Enumerators also faced political risks, as their efforts sought to influence political processes relating to land and housing allocation, two highly contentious issues in Tanzania.⁽²²⁾ For example, enumerators expressed concern that participation would affect the provision of public services or would result in intimidation from representatives of the ruling party. Because a single political party has held power since Independence, Tanzania has limited experience with protest. Where opportunities for complaint and protest have existed, these have generally been funnelled through the party system, thus limiting people’s experience of taking independent, unauthorized action to address their concerns.⁽²³⁾ Hyden states: “… few Tanzanians engage in collective action in order to promote or defend a particular idea or cause.”⁽²⁴⁾ In this context, the enumeration was highly visible and all the more politically risky. The risks were enhanced because the agencies overseeing land administration and allocation in the country “… suffer from malpractices such as corruption, nepotism and favouritism.”⁽²⁵⁾

Tarrow defines a social movement as “… collective challenges, based on common purposes and social solidarities, in sustained interaction with elites, opponents and authorities.”⁽²⁶⁾ By engaging in a sustained, public enumeration effort around the Kurasini eviction and in demanding land for resettlement, TFUP members made a strong collective claim on the Tanzanian state. While TFUP may be smaller in scale and scope than the kinds of movements that have dominated social movement scholarship in Europe and the US, it is nonetheless representative of the increasingly wide range of relatively small, yet transnationally connected movements taking root across the developing world.⁽²⁷⁾

The three-week enumeration conducted in Kurasini in August and September 2007 resulted in more than 2,500 households being interviewed and mapped by a team of 41 TFUP members.⁽²⁸⁾ Soon after the enumeration, in early October 2007, both owners and renters in Kurasini were given one month to leave their homes in preparation for large-scale demolitions. Owners of houses – who accounted for approximately 40 per cent of the population – were told by government officials that they would eventually be compensated for lost property; however, they were not told the precise compensation amounts until immediately prior to eviction, after the mobilization under investigation here had concluded.⁽²⁹⁾ In keeping with national policy, renters received no compensation.⁽³⁰⁾ For this reason, they were the focus of TFUP’s mobilization activities in Kurasini. Organizers stated that renters would have priority in receiving housing when land for resettlement was allocated.⁽³¹⁾ Beyond the limited compensation promised to owners, evictees received no group relocation assistance and the resettlement process was the responsibility of individual residents themselves.

III. Hypotheses

This section summarizes the paper’s formal hypotheses. The dependent variable for each is social movement participation, as measured by a yes/no variable indicating participation as an enumerator. As mentioned, since participation as an enumerator was physically and politically risky and time consuming, it was expected to reflect a strong commitment to being a participant in the TFUP movement.

a. Hypothesis 1: Payoffs

“If individuals receive payoffs, they will be more likely to participate as enumerators.”

From a rational choice perspective, people who receive individual payoffs from their involvement in a social movement will be more likely to participate. Selective benefits and individual payoffs provide incentives to contribute to the provision of public goods, where the free rider problem otherwise militates against participation.⁽³²⁾ In Kurasini, TFUP organizers believed that material benefits from participation as an enumerator would flow towards renters, and that this would motivate their participation. TFUP presented the enumeration and land allocation efforts as benefiting renters and, as such, owners could have only a limited expectation of payoffs from participation. TFUP organizers often spoke of their grassroots group of renter enumerators. Building on these observations, the expectation of this first hypothesis was that renters would dominate the pool of enumerators because they alone could expect significant benefits from serving as enumerators. The measure for the independent variable was self-reported owner/renter status as assessed through an open-ended interview question.

b. Hypothesis 2: Identity

“If individuals have greater identification with the TFUP movement, they will be more likely to participate as enumerators.”

Group identification is held to be a major motivation for movement participation.⁽³³⁾ This hypothesis builds on broad evidence that identity influences behaviour,⁽³⁴⁾ and on theoretical claims that the more an individual identifies with a social movement, the more likely they will be to participate in social movement activities.⁽³⁵⁾ Previous research also reveals that identity is not fixed and that individual and group identities grow out of experiences and relationships.⁽³⁶⁾ Individuals are also thought to possess multiple identities that exist in a hierarchy of salience.⁽³⁷⁾ This second hypothesis holds that individuals who report the TFUP movement identity as being more important than other dimensions of their personal identity will be more likely to participate as enumerators. The independent variable measure was assessed by asking interviewees to report their most important source of identity.⁽³⁸⁾

c. Hypothesis 3: Connection to place

“If individuals have a stronger connection to place, they will be more likely to participate as enumerators.”

Research shows that connection to place has strong behavioural implications.⁽³⁹⁾ Individuals with a strong connection to place feel more invested in its future and the public good and, as a result, are more likely to contribute to collective action.⁽⁴⁰⁾ Greater connection to place is also thought to enhance one’s sense of empowerment, because this is a product not only of relationships between community members but also between community members and their physical environment.⁽⁴¹⁾ Tuan⁽⁴²⁾ and Frederickson and Anderson⁽⁴³⁾ show that it is through interaction with the particulars of a location that one endows it with value and develops a connection to place. Following this reasoning, connection to place might be assumed to increase over time; therefore, duration of residence in one’s settlement was used as a measure of the independent variable, connection to place. A second measure was whether residents consider their settlement to be their “home”, irrespective of duration of residence. Both of these measures were employed and were assessed through open-ended questions soliciting interviewees’ duration of residence in their settlement (in months) and whether the interviewee considered their settlement to be “home” (yes/no).

IV. Data Collection and Descriptive Statistics

Data were collected through interviews with TFUP members between July and November 2007. Of the 81 interviews that were conducted, 41 were with volunteer TFUP enumerators and 40 were with TFUP members who did not participate as enumerators. This interview selection strategy allowed for a comparison of responses between TFUP members who did and did not participate. The sample included all individuals participating as enumerators. Non-enumerator interviewees were selected by distributing a request to all members of TFUP savings groups in Kurasini to participate in interviews and then selecting those respondents living closest to a central, randomly selected interview site. Interviews lasted approximately two hours and, although structured around a common set of questions, were open-ended. Interviewees were not paid but were compensated for travel to the interview venue if coming from elsewhere on that day.

Table 1 presents descriptive statistics for the study sample, including the variables associated with each formal hypothesis as well as several other key characteristics, including duration of TFUP membership, gender, age and religion. There were no statistically significant relationships between these additional characteristics and the likelihood of participation as an enumerator. With respect to the specific finding that gender was not statistically associated with participation as an enumerator, this does not mean that women do not play a large role in the TFUP movement or informal settlements more broadly. But with respect to the narrower question we posed in this paper – relating to the factors motivating participation as an enumerator – it appears that gender is not a major determinant of participation. In fact, as the formal hypothesis tests and qualitative evidence gathered through interviews show, there is compelling evidence that cleavages other than gender were significantly associated with and responsible for observed patterns of participation.

Table 1.

Descriptive statistics for the full study sample, including variables associated with the three formal hypotheses and other key sample characteristics

Continuous variables	n	Min	Max	Mean	SE mean
Months of residence in home settlement	81	6.0	672.0	222.8	15.7
Months of TFUP membership	81	0.1	72.0	14.0	1.7
Age (years)	81	22.0	71.0	43.3	1.2
Binary variables (n)	Category 1 (%)			Category 2 (%)
Ownership status (81)	Owners (71.6)			Renters (28.4)
Movement identification (60)	Primary identity is TFUP (66.7)			Primary identity is other (33.3)
Consider settlement “home” (81)	Yes (55.6)			No (44.4)
Gender* (81)	Female (64.2)			Male (35.8)
Religion (81)	Christian (45.7)			Muslim (54.3)

*

The dominance of female evictees likely reflects that the TFUP movement began as a women’s organization. While membership is open to men, the higher proportion of women in the movement reflects these origins.

V. Hypothesis Test Results

With respect to the first hypothesis, statistically significant results of a chi-squared test (Table 2) show that, contrary to expectations, property owners were more likely to participate as enumerators than renters, the group expected by TFUP to benefit from higher payoffs (p = 0.000). Owners accounted for 37 of the 41 enumerators (90 per cent). Also contrary to expectations, individuals who identified primarily with the TFUP movement and its affiliated savings groups were significantly less likely to participate as enumerators (chi-squared test: p = 0.000).

Table 2.

Results of statistical tests of formal hypotheses

Hypothesis measure	Enumerators	Non-enumerators	Difference (se) 1	Test
H1: Ownership (%) 2	90.2	52.5	37.7***	Chi-square
H2: Primary identity: TFUP (%) 3	36.6	97.5	60.9***	Chi-square
H3a: Mean months of residence in home settlement	282.2	161.9	120.3***(28.3)	T-test
H3b: Report settlement is “home” (%) 4	70.7	40.0	30.7**	Chi-square

NOTE: ***Significant at the 99.9% level, with p < 0.001; **Significant at the 99% level, with p < 0.01. n=81, including 41 enumerators and 40 non-enumerators. ¹Standard errors of the difference between means are reported for T-tests only. ²Measure is ownership status, assessed as owner versus renter. ³Measure is primary identity, assessed as TFUP versus “other”. ⁴Measure is whether interviewees report that their settlement is “home”, assessed in terms of yes versus no.

In keeping with expectations, individuals with stronger connection to their local settlement were significantly more likely to participate as enumerators. As reported in Table 2, for all interviewees, individuals who had lived in their local settlement for longer (Student’s T-test: p = 0.000) and those who considered it “home” (chi-squared test: p = 0.005) were more likely to participate as enumerators. Among the independent variables from all three hypotheses, property ownership was associated (at statistically significant levels) with measures of both connection to place and movement identification. Owners differed from renters not only in being significantly more likely to participate as enumerators but also in having significantly greater connection to place, according to both measures, and lower movement identification.⁽⁴⁴⁾

VI. Qualitative Analysis

The qualitative analysis builds on the three key relationships identified in the hypothesis test results: between participation and property ownership; participation and connection to place; and participation and movement identification.

a. Participation and property ownership

Confirming the outcomes of the formal hypothesis tests, interviewees’ narratives identified ownership as an important factor motivating participation as an enumerator. Interviewees also noted that ownership is an important feature of slum life more generally. As one renter in Kurasini commented: “Renters and owners think of their relationship with property, land and housing very differently.” In the course of fieldwork, as it became clear that more owners than renters were participating as enumerators, interviewees were asked why this might be. In total, 42 interviewees (including 23 owners and 19 renters) were asked why owners and renters might participate as enumerators at different rates. Responses were coded into nine mutually exclusive categories (Table 3). The most frequent responses (cited by more than 15 per cent of interviewees) are addressed in greater detail in the analytic narrative developed over the following sections.

Table 3.

Reasons provided by interviewees for the more frequent participation of owners as TFUP enumerators. Values are number and percentage of interviewees citing each reason (n=42)

Coded reasons for the more frequent participation of owners	Number of interviewees citing (n)	Percentage of interviewees citing (%)
Owners get payoffs from compensation	13	31.0
Owners believe in efficacy of their actions	10	23.8
Owners have more responsibility to their settlement	9	21.4
Renters are transient and “unimportant”	8	19.1
Owners have more free time	8	19.1
Renters don’t understand these issues	6	14.3
Owners have an interest in property	5	11.9
Poor information flow	5	11.9
Accidental / chance	5	11.9

i. Differences in free time and owner/renter participation

The issue of free time merits closer study because it was not considered in the formal hypotheses nor is it supported as an explanation by other narrative evidence. Although free time was mentioned as a factor by eight of the 42 individuals questioned, interviewees’ accounts nonetheless suggest this explanation is unlikely to explain the greater frequency of owner participation. For instance, several renters noted explicitly that: “…it is untrue that renters have less time to participate.” And several renters strongly argued that the perception that owners have more free time is a by-product of the large “cultural” differences that separate owners and renters.

Owners were thought by many interviewees to have higher incomes than renters, and as a result were also thought to have a less urgent need to work. A Student’s T-test showed, however, that there was no significant difference in the monthly household incomes of owners and renters (p = 0.732). From an income standpoint, owners were no less likely than renters to suffer financially should they participate as enumerators; the opportunity cost of participation was no different for the two groups. Furthermore, all owners were found to have occupations outside the home. Owners did not derive their full incomes from being landlords and, like renters, found it necessary to earn money in other ways. These findings, together with the relatively low priority given this explanation by interviewees, strongly suggest that factors other than free time are likely to account for the different rates of participation by owners and renters.

ii. Payoffs favour participation by owners

Thirty-one percent of interviewees (six renters and five owners) felt that payoffs from participation favoured owners. While these accounts confirmed that payoffs are important, as proposed in the paper’s first hypothesis, they also indicated that payoffs did not favour renters as expected.

The chain of reasoning that saw renters most likely to benefit from participation was based on the logic of TFUP organizers themselves. They felt that since the potential benefits of a successful enumeration outcome (i.e. reaching TFUP’s goal of a government allocation of land for resettlement) were to be targeted in the first instance towards renters, they would be more likely to participate. According to interviewees, however, because the government’s limited compensation package – offered only to owners – was based on the dimensions of an individual’s property, owners had a stake in the accuracy of the enumeration. As one renter remarked: “Owners participate most because they are the people who benefit most. Owners are responsible for the value of their house and the enumeration secures value.” Likewise, one owner commented: “Owners participate because they want to ensure the correct measurements are taken and this will … affect compensation values.”

Since compensation funds were derived from a single common fund, it was necessary not only to ensure that one’s own plot was measured accurately but that others’ were as well. To be certain of the enumeration’s accuracy, an owner would have to participate individually. Renters’ purported payoffs, in contrast, centred simply on seeing that an enumeration took place and that government responded to lobbying by TFUP leaders with a grant of land. Individual participation was not necessary to achieve this collective outcome.

The payoffs for owners were also shorter term than those expected to benefit renters, and were therefore more likely to motivate participation. Renters were expected to work towards an altogether new outcome, a grant of land from government, and would have to wait for the final result of the post-enumeration lobbying process to determine if their efforts were successful. Owners hoped to ensure that existing compensation processes were accurate. Their payoffs were not only shorter term but, being less politically contingent, were also easier to achieve.

iii. Efficacy of action: an issue of voice

The second most commonly cited reason for the more frequent participation of owners was greater belief by owners in their efficacy of action. Renters’ lower belief in their efficacy was listed by 10 of the 42 interviewees questioned (24 per cent), including five owners and five renters, as an important reason for lower rates of renter participation. An additional eight individuals (19 per cent), including five owners and three renters, stated – while not mentioning efficacy directly – that renters were transient and therefore felt that they were less “important”; whether directly or indirectly, 18 interviewees (43 per cent) raised the issue of efficacy as a reason for the lower rate of renter participation.

Interviewees frequently remarked that owners were more likely to “…expect a response from government.” For instance, one renter remarked: “Tenants are worried about whether their voices will be heard so are unwilling to put in time to local efforts.” Interviewees also held that owners were listened to more frequently, and that the promise of compensation indicated that government took their claims seriously. As one renter said: “The priority is given to owners, this is seen in the compensation plan.” One owner expanded on these sentiments, stating:

“Most owners participate because they have confidence. And their confidence comes from their property. Renters lack this confidence – confidence that their actions will bring returns and benefits – because they lack property. This is because renters are rarely considered and don’t benefit from decisions.”

These accounts parallel the findings of Cadstedt’s study of renter politics in Mwanza, Tanzania, which showed that renters were unlikely to register their complaints and take action on grievances.⁽⁴⁵⁾ She positions these findings in the context of Hirschman’s exit–voice–loyalty model, which provides a useful framework for understanding why renters were unlikely to participate as enumerators.⁽⁴⁶⁾ Hirschman argues that there are two avenues whereby people may address the declining performance of a firm, organization or state. The first is by abandoning it (“exit”); and the second by acting upon concerns to promote change (“voice”). Voice, he states, includes “…any attempt to change, rather than to escape from, an objectionable state of affairs”.⁽⁴⁷⁾

In Tanzania, as in many other one-party states, there is a limited tradition, and little personal experience, of voicing complaint against the government.⁽⁴⁸⁾ As Hyden notes, social movements have been largely absent from Tanzania.⁽⁴⁹⁾ This relative lack of mobilization experience raises the effective cost of pursuing voice in Tanzania.⁽⁵⁰⁾ In general terms, Hirschman argues that the likelihood of people voicing complaints and acting on concerns depends on whether previous attempts have met with success.⁽⁵¹⁾ This reasoning parallels the conviction of many interviewees in this study: owners had greater belief in the efficacy of their action because they had had a positive past experience with government responsiveness to their claims in the form of the limited compensation offered to owners.

Comparing exit and voice, Hirschman contends that exit is generally less costly to pursue. However, he also notes that while exit is common in the economic sphere, voice is more frequently observed in politics. For the poor, and in circumstances where the state is responsible for the adverse conditions faced by individuals – as is the case with slum evictions in Kurasini – options for exit are radically reduced. Since the slum dweller condition is defined, in part, by an inability to exit (in general, people would not live in slums if they could leave), opportunities for choosing exit in Kurasini were additionally constrained.

Having to face high costs of voice (relative to the perceived benefits of participation) and with exit effectively unavailable, what options were actually open to renters prior to eviction? Their situation can be understood through a conceptual addition to the exit–voice–loyalty model proposed by Tung.⁽⁵²⁾ Tung dubbed situations where voice and exit both appear unavailable as “autism”.⁽⁵³⁾ This condition is realized when there are few or no avenues for exit and voice is seen as too costly. According to Tung, in these circumstances people adjust their behaviour to fit their situation rather than try, against extreme odds, to change it.⁽⁵⁴⁾ The result is a state of deep resignation. According to interviewees, resignation of the kind described by Tung characterized the renter experience in Kurasini. As one renter concluded: “Tenants count for nothing in the whole process, so people don’t include them. It’s assumed they’re unimportant to these issues – by the city, owners and tenants.”

b. Participation and connection to place

i. Connection to place and “responsibility”

Connection to place emerges from the hypothesis test results and from the accounts of interviewees as a potentially important motivation for participation as an enumerator. Interviewee narratives suggest that connection to place may be closely related to an additional concept – responsibility to place – held by many to be important to participation. Nine interviewees (21 per cent), including five owners and four renters, remarked that owners have a greater sense of responsibility to their settlement.

In their comments, interviewees frequently tied together the issues of connection and responsibility to place. As mentioned by one renter: “Renters are seen, and see themselves, as temporary residents with less responsibility.” Likewise, one owner stated: “Tenants are just staying temporarily and this has implications for their responsibility to the community.” Another owner commented: “Tenants think they will leave at any time. They are, therefore, less concerned with the community.” These remarks substantiate the hypothesis tests results, which showed that duration of residence in one’s settlement and the likelihood of considering it “home” were both significantly and positively related to the likelihood of participation. They also extend the quantitative results by suggesting that responsibility to place plays an important mediating role between connection to place and participation.

ii. Connection to place and perceptions of challenges facing community

To further examine the relationship between connection to place and participation, interviewees living in the first area of Kurasini to be evicted (n = 66) were asked which challenges they felt were most important in the ward. Their responses were coded into several mutually exclusive categories, which were further broken down into two broad categories. “Long-term challenges” were those that confronted the settlement over a multi-year period, and “short-term challenges” were more immediate, and in all cases were associated with the impending evictions and demolition of Kurasini. Owners and renters differed significantly in their perceptions of the most important challenges facing Kurasini. Seventy-six per cent of owners reported long-term challenges as being most important, compared to only 38 per cent of renters, a significant difference (p = 0.003). Renters typically listed the eviction itself as the most pressing challenge facing their settlement.

Capturing renters’ sentiments more generally, one renter noted: “The main problem is the threat of eviction and resettlement. Other problems don’t matter anymore.” Owners, on the other hand, described such long-term challenges as water supply, sanitation, transportation, lack of a dispensary and the presence of unsafe industry. As one owner stated: “The main problem is water supply. There is also flooding during rains, waste disposal problems and road construction has affected our water.” This difference dovetails with the accounts of interviewees that owners have greater connection and responsibility to place, which in part accounts for their participation. However, this is also a challenging finding since, while it may often work to the advantage of owners to view their settlement’s challenges over the long term (in terms of advancing their investment in property, for example), this dimension of connection to place appears counter-intuitive when the settlement itself has no long-term future.

iii. Connection to place and “loyalty”

The concept of “loyalty” from Hirschman’s exit–voice–loyalty model provides a useful theoretical framework for understanding how connection to place translates into participation as an enumerator.⁽⁵⁵⁾ Hirschman defines loyalty as a “…special attachment to an organization” and notes that when people are more “loyal” they are more likely to choose voice, even when the costs are high.⁽⁵⁶⁾ In a spatial context, this concept of loyalty is similar to the idea of connection to place. One of this paper’s core hypotheses contended that individuals with greater connection to place would be more willing to engage in costly social movement activities. This parallels the logic developed by Hirschman regarding loyalty, which holds that loyalty increases the likelihood of pursuing voice by effectively reducing the perceived costs of this course of action.

c. Participation and movement identification

Contrary to expectations, individuals who identified primarily with the TFUP movement were significantly less likely to participate as enumerators than those who did not (28 per cent participation versus 96 per cent). In particular, it was renters – already shown to account for a small fraction of enumerators – who identified primarily with the movement. When asked the open-ended question, “What is your most important identity?” 96 per cent of renters reported they identified primarily with the movement, compared with 55 per cent of owners.

To delve into the complex relationship between movement identification and participation, supplementary interviews were conducted with a range of Tanzanian experts on slum dweller politics. These 20 interviewees included academics, civil society leaders and public officials, all of whom were familiar with TFUP and its enumeration effort in Kurasini. Their responses suggested that renters’ strong movement identification did not reflect deeper commitment to the movement. Instead, it was a product of “desperation”. These expert interviewees contended that the very poor often joined movements such as TFUP in the hope of escaping their economic and social situation. They also argued that, in the context of African informal settlements, renters were often more desperate than owners and were therefore more eager to cling to an identity that offered some hope of improving their quality of life.

A professor in the Department of Sociology at the University of Dar es Salaam stated: “These groups are a lifeline for the poor … their desperation and vulnerability [referring to renters] makes them eager [to join movements like TFUP].”⁽⁵⁷⁾ Maria Saguti Marealle, Director of UN−Habitat’s Cities Alliance programme in Dar es Salaam, reiterated the connection between renters’ desperation and their desire to identify with movements such as TFUP, saying: “Renters joining savings groups feel as though they are valued, despite their dire situation. They get personal value from their involvement.”⁽⁵⁸⁾ The desperation of renters was further characterized by the country director of UN−Habitat, Philemon Mutashubirwa, who remarked:

“With renters, there is … a culture of distrust and desperation. They haven’t been compensated so don’t really want to waste any time contributing to any effort that might not yield benefits.”⁽⁵⁹⁾

The dominant theme in these comments is that renters are desperate to escape their situation, and as a result are more willing to adopt identities that may appear as a lifeline. The previous sections of this paper – outlining renters’ limited connection to place and low belief in efficacy of action – give a sense of why renters might have a greater sense of desperation than owners. Furthermore, a common theme in the narratives of TFUP members is a feeling of renter helplessness. Describing this helplessness, one renter stated: “Renters are helpless. They can’t help themselves.” One owner agreed that renters must struggle with a feeling of inconsequentiality, saying: “Tenants’ voices will not be heard. Tenants don’t count for much in the community.”

The desperation to which these interviewees refer parallels Tung’s characterization of the resignation that results when neither exit nor voice are feasible options.⁽⁶⁰⁾ It also concurs with the image of renters emerging from the narrative developed here as individuals without choices. In terms of Hirschman’s exit−voice−loyalty model, renters were faced with both a limited ability to exit and high costs of pursuing voice.⁽⁶¹⁾ In this light, their greater willingness to adopt the movement identity is not surprising. As a result, however, their strong identification should not be seen as reflecting deeper commitment or willingness to act on behalf of the movement. Rather, identification with the movement reflects a coping strategy for individuals who find neither voice nor exit options realistically available. This finding has important implications for community leaders and development practitioners, who might otherwise expect that greater movement identification reflects higher levels of commitment and willingness to participate.

VII. Conclusions

The results of this research show that the majority of participants in TFUP enumerations in Kurasini were owners, not renters as expected by movement organizers. The narrative developed here indicates that three factors favoured the participation of owners over renters: the nature of payoffs from participation; owners’ greater belief in their efficacy of action; and owners’ greater connection to place. These findings have implications for the understanding of social movement participation by slum dwellers and for development policy.

While movement organizers expected renters to participate to secure a successful enumeration outcome (where they would receive plots through a government allocation of land), TFUP members argued that owners and renters faced different payoffs, and that owners’ payoffs were associated with ensuring the accuracy of the enumeration process. The payoffs cited by members as driving owner participation were more individualized, shorter term, less politically contingent and, ultimately, easier to realize than the payoffs believed by movement organizers to face renters. Interviewees’ narratives showed that owner participation was also motivated by greater belief in the efficacy of their action and greater connection to place. This last point was supported by quantitative results demonstrating that owners had lived in their settlement for significantly longer than renters, were more likely to consider it home and viewed community challenges over a longer time horizon.

The empirical results from Kurasini show a striking overlap with several bodies of theory, most notably Hirschman’s exit−voice−loyalty model.⁽⁶²⁾ In general terms, Hirschman argued that belief in the efficacy of one’s actions and “loyalty” (argued here to be akin to connection to place, in geographical terms) reduce the perceived cost of pursuing voice. In this case, owners, who had greater belief in their efficacy of action and greater connection to place, participated more frequently. An extension to Hirschman’s model, developed by Tung, contributes further to an understanding of renter behaviour.⁽⁶³⁾ Renters, who were significantly more likely than owners to identity with TFUP, found both exit and voice unavailable. Tung argued that people in situations where both of these options are unavailable are characterized by a sense of resignation and powerlessness, precisely the characteristics attributed to renters by interviewed TFUP members.

This research offers several policy lessons. The first is that even in seemingly cohesive groups such as the members of TFUP, there may exist powerful cleavages capable of significantly influencing people’s willingness to participate in collective action. The most important cleavage among TFUP members was between owners and renters, who differed in almost all aspects of participatory behaviour. The discovery that TFUP mobilized a markedly different group than intended reinforces a second lesson: while payoffs are not the only factor motivating slum dweller mobilization, they are critical to shaping behaviour. The results demonstrate that it is important for movement organizers to understand which kinds of payoffs are likely to influence participation. The paper’s findings concerning owner and renter conceptions of the time horizons of local challenges suggest that renters may be unlikely to participate in activities focused on the long-term future of informal settlements, and that more attention needs to be given to the psychological dimensions of being a renter. Indeed, renters’ limited belief in their efficacy of action and the opinion, held by renters and owners, that renters’ views were of little value, were important factors in the non-participation of renters. The final lesson is that strong identification with a social movement, such as TFUP, may not translate into movement participation. TFUP, with its promise of change under desperate circumstances, was easy for renters to identify with, but that identification did not translate into participation in collective action in the face of renters’ highly uncertain payoffs, limited connection to place and a belief that their voices did not matter.