Targeted conditional cash transfer programmes linked to human development objectives started in the 1990s in Latin America and have spread worldwide, having been adopted in 64 countries. While the targeting dimension of these programmes has become increasingly more accepted in different policy, practitioner and academic circles, their conditional component still elicits significant controversy. What are the independent effects of conditionalities, beyond the income effect of cash transfers?

On the one hand, arguments in favour of conditionalities maintain that they can rectify market failures such as a lack of information, high intertemporal discount rates and imbalanced intrafamily bargaining power that would prevent families from making optimal investments in the education of their children. Conditionalities can also increase private investment in education, which may be below the social optimum due to the existence of positive externalities. Finally, they also serve to legitimise and justify, at the political economy level, government transfers being disbursed to beneficiaries.

On the other hand, arguments against the inclusion of conditionalities state that access to a minimum income is a basic human right, and thus should not be conditional on certain behaviours. Another argument emphasises that labelling programmes as child allowances would produce an effect similar to conditionalities, by ensuring investments in the health and education of beneficiary children. Finally, there is the idea that conditionalities could have negative effects through the stigmatisation of beneficiaries and, potentially, the exclusion of the most vulnerable from social programmes, since they are less likely to comply with conditionalities.

The available evidence is slightly in favour of the existence of impacts of conditionalities beyond the effect of the cash transfer component, particularly with regard to educational impacts. However, they have been inconclusive so far. In that regard, Baird et al. (2013), in their systematic review of 35 studies, suggest that the level of enforcement associated with the monitoring of conditionalities is the main channel through which conditionalities would have an independent and additional impact on educational outcomes.

Paiva et al. (2016) look at this independent effect of conditionalities in the context of the implementation of the Bolsa Famìlia programme in Brazil. Given that the programme’s coverage and its rate of conditionality monitoring are not correlated at the municipal level, a growth curve model (Singer and Willet 2003) is used to measure the independent impact of the conditionality monitoring and level of coverage at the municipal level on educational outcomes—namely, drop-out and progression rates—controlling for confounding variables, in a context whereby these rates have clear descending and ascending trajectories, respectively. The independent variables of interest are programme coverage (which was assumed to be a proxy for its cash transfer component) and the rate of school attendance monitoring for basic education—the first nine years of schooling (which was assumed to be a good proxy to measure the conditionality component)—both at the municipal level.

The results of the growth curve models do not suggest any statistically significant association between the coverage of the Bolsa Família programme and educational outcomes. However, the variable representing conditionalities (school attendance monitoring) had a positive effect on the outcomes of interest: the greater the monitoring, the lower the drop-out rate, and the higher the school progression. The growth curve model also allowed us to assess whether the variable of interest had any impact on the evolution of the educational outcomes between 2008 and 2012. The association between conditionality monitoring and educational outcomes found for the initial status is not found for the trajectory of the outcomes. There is a clear convergence between municipalities towards lower drop-out rates and higher progression rates. This trend suggests that, despite the positive effect of conditionalities, the most important factor driving the progression of the two indicators is the convergence trend, which actually reduces the space for a sizeable impact of both cash transfers and conditionality monitoring at least for the basic level of education.

Based on previous studies, these findings could be considered somewhat unexpected. However, there are peculiarities to the Brazilian context that might help to explain them. Brazil is a middle-income country with a strong supply of public education. While its quality definitely continues to be an issue, only a very small part of the school-aged population do not have access to public education. Problems that could potentially affect school attendance have been addressed through different programmes, such as the National School Feeding Programme (Programa Nacional de Alimentação Escolar—PNAE) and the National School Transportation Programme, both with national coverage.

In such a context, it is not surprising that a relatively small cash transfer only has a limited (if any) effect on educational indicators. However, as this transfer may represent the only stable source of income for the family and is conditional on school attendance, it may still have some small but statistically significant effect on these indicators. Impacts on secondary education may be larger and will be considered in another study.

This One Pager n is part of the UK Department for International Development (DFID) supported project: “Brazil & Africa: fighting poverty and empowering women via South-South Cooperation., and was first published on June 13, 2016.

References:

Baird, S. et al. 2013. “Relative Effectiveness of Conditional and Unconditional Cash Transfers for Schooling Outcomes in Developing Countries: A Systematic Review.” Campbell Systematic Reviews, No. 8.

Paiva, L.H. et al. 2016. “The effect of conditionality monitoring on educational outcomes: evidence from Brazil’s conditional cash transfer programme.” IPC-IG Working Paper 144. Brasília: International Policy Centre for Inclusive Growth.

Singer, J.D., and J.B. Willet. 2003. Applied Longitudinal Data Analysis: Modelling Change and Event Occurrence. Oxford: Oxford University Press.