From protection to production: productive impacts of the Malawi Social Cash Transfer scheme

Abstract

The Malawi Social Cash Transfer (SCT) scheme is part of a wave of social protection programmes providing cash to poor households in order to reduce poverty and hunger and promote child education and health. This paper looks beyond the protective function of such programmes, analysing their productive impacts. Taking advantage of an experimental impact evaluation design, we find the SCT generates agricultural asset investments, reduces adult participation in low skilled labour, and limits child labour outside the home while increasing child involvement in household farm activities. The paper dispels the notion that cash support to ultra poor households in Malawi is charity or welfare, and provides evidence of its economic development impacts.

Disclaimer: The views expressed in the Work are those of the Author(s) and do not necessarily reflect the views of the Food and Agriculture Organization of the United Nations.

Keywords:

• cash transfers
• impact evaluation
• income-generating activities
• Malawi

Notes

Disclaimer: The views expressed in the Work are those of the Author(s) and do not necessarily reflect the views of the Food and Agriculture Organization of the United Nations.

1. For additional information on the SCT scheme and its context, see: http://www.cpc.unc.edu/projects/transfer

2. Miller et al. (2008a, 2008b, 2010) provide detailed explanations of the evaluation design and targeting procedure.

3. We ran t tests to see whether the characteristics of attrition and panel households were statistically different. These tests revealed that nearly all factors were statistically the same across attrited and panel households. The same result was found when running a probit on attrition status. However, programme participation was statistically significant in both the t tests and the probit, indicating that treatment households were 5.5 per cent more likely to drop out than controls. It is probable that this result is driven by the 17 households found to be ineligible. Due to the small number of observations, we were unable to further statistically analyse the systematic nature of ineligibility or deaths in the original sample.

4. Ganyu labour is a type of low wage casual labour performed in Malawi. More detail on this type of labour is provided in Section 2.3 and in Dercon (2002) and Whiteside (2000).

5. Although we follow the methodology described by Miller et al. (2011), for the purposes of not losing observations in the empirical analysis we assign the arbitrary value of 0.1 in the denominator when calculating the dependency ratio for households with no able-bodied adults to assure that its value for those households is greater than three.

6. Propensity scores are simply the estimated probability of eligibility given a series of explanatory variables that are anticipated to be linked to participation.

7. We also considered applying propensity scores used as weights in an inverse propensity weighted system but found it was not as effective as a standard PSM approach.

8. Child level regressions also include the age, gender and orphan status of the child.

9. We also tested the robustness of our results with nearest-neighbour matching with replacement, using caliper widths of 0.1 and 0.5, and bootstrapping standard errors with 50 repetitions. With few exceptions we obtained similar outcomes.

10. Angrist and Pischke (2009) explain that a standard regression model can correctly estimate treatment effects across experimental groups, even for limited dependent variables.

11. The category ‘Leisure’ was not specifically defined in the survey instrument. We assume it refers to non-labour, non-schooling activities.

12. In order to check the robustness of our results with respect to household labour constraints, we also ran regressions dividing the sample according to the household dependency ratio, creating a labour-unconstrained group for households with a dependency ratio below three, and a labour-constrained group for those households with the dependency equal to or greater than three or for which the dependency ratio could not be estimated due to there being no able-bodied adults in the household. The findings for these regressions are, with few exceptions, consistent with the findings when households are divided according to having none or at least one able-bodied adult member.

13. Although we do not provide a full assessment of this literature in this paper, Deere and Doss (2006), Peterman et al. (2010) and Quisumbing et al. (2011) review the evidence on gender and asset ownership, Quisumbing et al. (2011) review the gender dimension of response to shocks, while FAO et al. (2010), FAO (2011), and World Bank (2011) present thorough reviews of gender in terms of agriculture and development.