An experimental analysis of network and group formation for collective action


ESRC logo
Funded by ESRC
DIFD logo
Funded by DFID

Zimbabwe’s first resettlement programme created inclusive communities

Photo: A villager making a risky decision in the group formation experiment
A villager making a decision
in the group formation field experiment

Zimbabwe’s first land reform and resettlement programme in the early 1980s led to the formation of highly inclusive communities. This is the conclusion from a study of 19 villages created under the programme. The poor are not excluded when, within these villages, people set up groups in order to address shared problems or take advantage of collective opportunities. Further, while men and women do reveal a tendency to separate into single sex groups, this is not owing to a lack of trust between the sexes and, possibly for that reason, female headed households are not excluded from community based organizations (CBOs). Family, clan, and religion all play important roles in bringing neighbours together and providing a basis for the trust required for them to act collectively. It remains to be seen whether the current land reform and resettlement programmes will do as well along this important dimension.

Implications for practitioners

Photo: Villagers constructing a map of kinship ties
Villagers constructing a map of
family ties

These findings will come as good news to development practitioners who, as part of community based development initiatives, engage existing CBOs or encourage people to form new groups and organizations in order to take part in interventions. However, it is important to bear in mind that this study focuses on a particular type of village.

That the tendency for men and women to separate into single sex groups is not owing to a lack of trust between the sexes and that religions support trust between the sexes and provide a basis for informal enforcement will be of particular interest to practitioners in the African microfinance sector.

Project design for “An experimental analysis of network and group formation for collective action”

Photo: A villager making a risky decision in the group formation experiment
A village research workshop in progress

This study used data from a quasi-experiment and a field experiment to investigate group formation for collective action in 19 resettled Zimbabwean. The quasi-experiment occurred as a result of actions taken by the Zimbabwean government. In the early 1980s, displaced people were resettled in new villages made up of unrelated and often unacquainted households. In order to survive and prosper, these villagers had to solve various problems of collective action. To varying degrees, the villages in the study did this by forming CBOs. The field experiment, conducted in the same resettled villages, involved a game that mimicked situations in which development agencies, non-governmental organizations, or government bodies invite villagers to form groups rapidly in order to address a shared problem or take advantage of a collective opportunity.

Analyses of the data on the groups formed within the field experiment and the CBOs formed in the new villages in the early 1980s and sustained throughout the following two decades with data on village geographies, kinship networks and household characteristics we investigated who groups and who groups with whom revealed a number of findings, details of which are presented below.

Researchers: Abigail Barr, Marleen Dekker, and Marcel Fafchamps

Findings in more detail

Photo: Villagers constructing a map of kinship ties
A field researcher explains the group
formation game to villagers

The poor are not excluded: The analyses revealed no evidence of the poor being excluded or choosing to exclude themselves from group formation and no evidence of the rich grouping with the rich and the poor with the poor. Indeed, the villages in which average wealth was lowest at the time of resettlement had the densest networks of CBO co-membership throughout the subsequent two decades. In these villages, it was the wealthier households who were instrumental in setting up the CBOs, possibly because, for them, land clearance and homestead-building was easier and faster. Initially, the poorer households would have had little time for anything but land clearance and building. However, by 1985 they appear to be just as engaged as the wealthier households.

Women and men do tend to separate into single sex groups but this is not owing to a lack of trust between the sexes: It will come as no surprise to anyone who has convened meetings in sub-Saharan Africa, that when women and men were invited to form groups quickly during the field experiment they tended to separate into male and female groups. However, this tendency was not owing to women finding it easier to trust other women and men other men. In fact, when trust was important there was less sex separation. Further analysis indicates that this is because trust is stronger within families and among people who belong to the same religious congregation and neither families nor religions are sex segregated.

Religions and families support the informal enforcement of collective agreements: As well as supporting the formation of mixed sex groups when mutual trust is important, families and religions provide a context within which people can hold each other to account. 

The relationships formed by belonging to the same CBO are valued: Co-members of CBOs were more likely to group together when invited to do so during the field experiment except when the agreements holding those groups together could be socially enforced. At first glance, this finding appears odd. However, it is consistent with community based organization co-members wishing to protect their on-going valuable relationships from the potential harm that could be done to them if one or other party was tempted to renege on an agreement within the experiment. When the grouping agreements were enforced by us there was no such risk. When they were supported by trust alone, they could be broken but no one would ever know. So on-going relations would be unharmed. To make informal social enforcement possible, we set the game up so that people who wished to renege on their grouping agreements had to tell everyone that that was what they were doing. So, a renegade could have upset on-going relationships and this represented a risk. Not grouping was the easiest way to avoid that risk. That CBO co-members did not group together under this last treatment is an indication that they value those co-memberships.

Academic Papers

The formation of community based organizations in sub-Saharan Africa: An analysis of a quasi-experiment” by Abigail Barr, Marleen Dekker, and Marcel Fafchamps. CSAE Working Paper WPS/2010-21.

Who Shares Risk with Whom under Different Enforcement Mechanisms?” by Abigail Barr, Marleen Dekker, and Marcel Fafchamps. BREAD Working Paper 267.

Bridging the gender divide: An experimental analysis of group formation in African villages” by Abigail Barr, Marleen Dekker, and Marcel Fafchamps. BREAD Working Paper 268.

Status and Egalitarianism in traditional communities: An analysis of funeral attendance in six Zimbabwean villages” by Abigail Barr and Mattea Stein. CSAE Working Paper WPS/2008-26.

Other resources

Scripts for the Field Experiment (PDF file)