News docente/cattedra

Brown Bag Lunch Workshop on Impact Evaluation

The Economic and Social Development Department (ES) of FAO is pleased to announce the following seminar:


Combining social protection with rural development interventions: non-experimental evidence from Lesotho


Presented by


Ms Noemi Pace, Economist, Social Policies and Rural Institutions (ESP)


Friday, 22 November 2019

14:00 - 15:30 hrs


Iran Room (B116bis)



In poor rural settings, characterized by inexistent or incomplete insurance and financial markets, individual risk preferences represent one of the channels driving the shift from low-return/low-risk activities towards high-return/high-risk activities. This study takes advantage of the data collection for the evaluation of the impact of an unconditional cash transfers program, the Child Grant Program (CGP), and a comprehensive community development package, the Sustainable Poverty Reduction through Income, Nutrition and Access to Government Services (SPRINGS), in rural Lesotho. The study has two main goals: 1) Investigate the CGP and CGP+SPRINGS effects on risk preferences: with this analysis we aim to estimate the link between inclusion in one of the three treatment arms, i.e. comparison group, exclusive participation into CGP, and participation in both CGP and SPRINGS, and our measures of risk preferences; 2) Investigate whether our measures of risk preferences are correlated with real-life risky choices, defined as farming and investment strategies with both a higher expected return and greater variance of returns. We implemented a mediation analysis which allows us to disentangle between the direct impact of the programmes on real-life risky choices and the indirect impact mediated through changes in risk preferences. From the analysis of the programmes’ impact on risk preferences we find that participation to CGP and SPRINGS affects both willingness to take risk and risk taking in laboratory experiments in the field. From the analysis of the link between risk preferences and real-life risky choices, we find that while our experimental measures are not correlated, risk preferences measured through survey data are good predictors of real-life risky choices.

Please contact Prof d'Errico if interested in participating.