Does the “Conditional” in Conditional Cash Transfer design improve program ability to reduce inter-generational poverty? Rachel Cassidy Friday 12th February, 2010 Rachel Cassidy, MSc Economics for Development Outline CCTs: a brief background Conditionality: the theory • Advantages • Disadvantages Conditionality: the empirics • Mexico‟s PROGRESA/Oportunidades • De Brauw and Hoddinott (2008) Rachel Cassidy, MSc Economics for Development CCTs: where it’s ATT Opportunity NYC PROGRESA PATH Cash For Education Eduque a la nina Red de Oportunidades Familias en Acción Bono de Desarollo Humano HIV/AIDS Scheme Bolsa Escola/Familia Jefes y Jefas de Hogares Chile Solidario Rachel Cassidy, MSc Economics for Development CCTs: an overview Cash transfer schemes; targeted at the poor; payment tied to compliance with health/education conditions. => Break intergenerational poverty transmission by encouraging human capital investment. “As close as you can come to a magic bullet in development.” Nancy Birdsall, president of the Centre for Global Development. Rachel Cassidy, MSc Economics for Development Why make transfers CONDITIONAL? Some economic theory… • Conditionality creates a PRICE EFFECT: • Brings marginal private benefit of health/education closer to marginal social benefit (by increasing benefit, or by reducing opportunity cost; whichever way you look at it!). • Pre-intervention MSB > MPB, because of POSITIVE EXTERNALITIES to human capital formation. • Transfer WITHOUT conditionality… => INCOME EFFECT only, => Increases in health/education very small. Rachel Cassidy, MSc Economics for Development More economic justifications for conditionality Why else might private decisions be suboptimal? • Poor households have imperfect information on the benefits of schooling/health. • High discount rate/credit constraints => earnings from child labour NOW more important than potentially higher lifetime earnings. • Discrimination against girls? • Household decision-making: bargaining position of mothers is weak? Rachel Cassidy, MSc Economics for Development And what about the politics? Reducing the stigma: a “contract”, not a handout. Social transfer systems receive more support if tied (as demonstrated by the widespread acclaim for CCT‟s). • Tangible targets for judging policy success. • “We know what‟s best for the poor”?.. Rachel Cassidy, MSc Economics for Development Drawbacks of conditionality Demeaning to the poor? Imposes costs on recipients… • Some can‟t meet the requirements. • “Morally atrocious”-social protection is a UN- declared human right that cannot be denied- Freelander (2007). If conditions don‟t align with poor‟s preferences, act as welfare-reducing constraints. Admin Costs: 2% of PROGRESA‟s annual budget… • 2% of $2.2bn = mucho dinero Rachel Cassidy, MSc Economics for Development So how big are the benefits? Some empirical evidence… Case study: Mexico‟s PROGRESA An economist‟s dream… A RANDOMIZED CONTROLLED TRIAL!!! Rachel Cassidy, MSc Economics for Development De Brauw and Hoddinott (2008) PROGRESA‟s transfers had 3 components: • Cash for food (the alimento); cash for school supplies; and cash conditional on attendance (the beca). Conditions for the beca: school attendance of >= 85%. Prove this with an E1 form, signed by the teacher. However, A SIGNIFICANT NUMBER OF HOUSEHOLDS (around 10%) NEVER RECEIVED THE FORM, YET STILL RECEIVED TRANSFERS. • No way of monitoring the child without the form, so these transfers UNCONDITIONAL. Rachel Cassidy, MSc Economics for Development A way to test the impact of conditionality… Confine sample to those who received the beca (with or without the E1 form). Regress enrolment on „treatment‟ (==1 if received the form; ==0 if not) and calculate ATT etc for conditionality. But what if households THOUGHT they were being monitored, even without the form? A more sophisticated test: • Using survey data, identify two groups: group A if received form AND UNDERSTOOD conditionality on schooling; group B if didn‟t receive form AND DIDN‟T understand conditions. • Group A definitely „treated‟ with conditionality. Group B definitely not. Rachel Cassidy, MSc Economics for Development The Econometrics Ensuring unconfoundedness: lots of x variables (child, household, and community characteristics). Probit model • p(enrolled) if treatment== 1 vs. 0 What if treatment still not fully random? • Nearest-neighbour matching. • Results almost identical to ordinary probit model. Rachel Cassidy, MSc Economics for Development Results ATT positive and significant: • Conditionality increases probability of enrolment. BUT, disaggregating by grade => treatment effect insignificant; except for grades 6, 7 and 8. Rachel Cassidy, MSc Economics for Development Results (continued) Rachel Cassidy, MSc Economics for Development Results (continued) Rachel Cassidy, MSc Economics for Development Conclusion Little benefit of conditioning transfers based on enrolment in primary school. Large benefit of conditioning transfers at entry into lower secondary school. => Considerable efficiency gains possible if CCTs calibrated more carefully. Rachel Cassidy, MSc Economics for Development Bibliography De Brauw and Hoddinott, Must conditional cash transfer programs be conditioned to be effective?, IFPRI (2008). De Janvry et al, Journal of Development Economics (2006). De Janvry and Sadoulet, Conditional Cash Transfer Programs: are they rally magic bullets?, GFAE (2004).