if group composition were the result of a ‘biased’ decision making processes, we should expect the
method of recruitment to be correlated with the degree of heterogeneity, which from table 7 is not.
5 Conclusions
Overall, the empirical evidence in this paper suggests that self-help groups are an important source
of income for certain categories of people, e.g. women. Individual earnings within the group seem
to increase with age more than with job experience (as proxied by the number of years spent in the
group), and are negatively correlated with the time spent in the current place of residence. This
probably reflects the precarious living conditions of the slums, which induce most skilled people to
leave if they have the opportunity to do so. Production groups also serve an insurance function by
giving access to informal credit: 64 percent of the respondents borrow from the group from individual
members in case of need. An important channel to gain access to such loans seems to be the ability
to speak the same language as the chairperson,leading to a probability increase of 20 to 25 percentage
points at the mean. Group composition affects both the extent to which borrowing can be carried
out within the group, and the organization of production in terms of division of labor, compensation
schemes, and sanctioning technology. The data at hand do not allow to estimate these effects
precisely, nor do they give insights into how groups get to be more or less heterogeneous. However,
they seem to suggest that fragmented groups may be limited in their choices by the need to curb
the higher potential for conflict within the group. To the extent that the resulting choices (lower
access to group loans, simpler production and compensation schemes, lower ability to sanction) are
suboptimal from an e¢ciency point of view, the case for understanding the roots of social cohesion
and ‘social capital’ within these groups is strengthened.
References
[1] Abraham, A., J-M. Baland, and J-P. Platteau (1998), “Groupes Informels de Solidarité dans un
Bidonville du Tiers-Monde: Le Cas de Kibera, Nairobi (Kenya)”, Non-Marchand, 2, 29-52.
19