Cash transfer interventions form increasingly important parts of social protection systems in most countries. Usually, cash transfers are evaluated against their first order effects on poverty or human capital, with their impact on social relations within and between households often relegated to discrete comments on „stigma‟, „resentment‟, and sharing, such as reduction of remittances and other support. Judgements on the desirability of cash transfers therefore normally attempt to balance the benefits for poverty reduction with any negative consequences on social relations. It is rare to find analyses of cash transfers that transcend this to comment on how cash transfer programmes impact on social relations and how social relations in turn impact on poverty. In other words it is rare to find comments on second order effects on poverty, accounting for social relations. In practice, negative consequences on social relations are often cast as problems to be solved within the framework of the intervention, rather than reasons to reconsider the intervention itself.

Using evidence from Oxford Policy Management‟s evaluations of cash transfer programmes in Kenya, Malawi and Zimbabwe, we suggest reconsidering this approach. We argue that the current evaluation designs, may first implicitly put a lower value on social relations than the value given by individuals in targeted areas. Second, it may understand poverty as a more individual construct than most individuals in targeted areas would understand it. Third, social relations have significant effects on livelihoods and poverty through shared inputs, mutual support, and pooled labour. These second-order effects may take longer to occur but are important. Evidence from our evaluations is provided to support these points.

We suggest conceptualising cash transfers as ongoing processes of intervention in a complex system of social relations. Individuals are embedded within institutions at several levels: households, communities and national political and social institutions. Resources, power, and knowledge flow through these institutions. Cash transfer interventions operate through and affect these flows at each stage: awareness-raising, targeting, payment, case management and monitoring and evaluation. We provide several examples of this from our case studies.

We conclude that the impact of cash transfers on social relations is large, often positive but in some cases negative. This matters not only at a local level, but also at a national level. Changes in local political economy (caused by any intervention) have wider political impacts. Moreover, the outcome of cash transfer experiments is not always positive. This implies that donors should engage more with domestic political groups in designing and implementing pilot projects, and should focus more on social relations in their evaluation designs.