1 – Charity hazard in climate disaster response: The effect of humanitarian cash transfers on demand for agricultural insurance
Julian Roeckert
Potsdam Institute for Climate Impact Research (PIK); Ruhr University Bochum; RWI – Leibniz Institute for Economic Research,
Abstract: Providing assistance in the context of weather disasters can have unintended negative consequences if it disincentivizes households to adapt to climate risks. This dilemma, known as charity hazard, is well justified by economic theory, yet empirical evidence remains scarce. Our paper analyses the effect of randomly distributed cash transfers during an extreme winter event on the demand for agricultural insurance among pastoralist households in Mongolia. We find that receiving the transfer does not crowd out the uptake of insurance covering weather risks in the subsequent year. Among households that have purchased insurance in the past, we find that the cash transfers significantly increase the probability of repurchasing insurance.
2 – Horticulture helps: How home garden interventions alleviate food insecurity in crises
Dorothee Weiffen
ISDC – International Security and Development Center; Zero Hunger Lab, Humboldt-Universität zu Berlin
Abstract: Home garden interventions have demonstrated potential in enhancing nutritional outcomes, primarily through providing healthy food, but indirectly building up livelihoods. However, evidence from home garden interventions from settings experiencing multiple crises remains absent. Given the unique challenges posed in poly-crises settings, assumptions derived from more stable contexts may not be applicable, leaving a significant gap in our comprehension of the effectiveness of home garden interventions. Furthermore, the disturbing impact of crisis context itself on interventions remains largely understudied. Lastly, the role of mechanisms that impact intervention outcomes in such settings is barely studied. Understanding the impact pathways of home garden interventions on nutritional outcomes within poly-crisis contexts is paramount for tailoring interventions to suit the context and optimizing the allocation of limited resources.
To fill this knowledge gap, we leverage a case study from an integrated home garden intervention in South Sudan. Our study is located in the surroundings of Juba, where a significant share of households have sought refuge following displacement caused by the civil war. Additionally, severe flooding has impacted the households in this area during our study period. Coupled with high levels of poverty and food insecurity, our study population is confronted with a complex array of challenges, constituting a poly-crisis scenario. With the aim of enhancing household nutrition, the home garden intervention encompassed the provision of agricultural plots, seeds, tools, training, and a cash transfer. This comprehensive support package was facilitated by Malteser International and specifically tailored to benefit displaced farmers. The study follows a quasi-experimental design. With three waves of survey data from before, during and after the complete intervention, we first conduct a differences-in-differences analysis to assess the average treatment effect of the intervention on nutritional outcomes. Then, we conduct a mediation analysis on potential mechanisms through which home garden interventions impact nutritional outcomes. Specifically, we test the access and availability of nutritious food, the creation of income and assets, behavioral change and the improvement of agricultural and nutritional knowledge. We enhance our sample balance using entropy balancing.
Our preliminary findings indicate an average increase in food consumption by 36% (p<0.01) according to the Food Consumption Score and an average decrease food insecurity as measured through the Household Food Insecurity Access Prevalence Scale by 14% (p<0.01). Interestingly, significant impacts on dietary diversity remain absent. We identify the improvement in nutritional and agricultural knowledge as well as the building of an asset stock as strong mechanisms for the treatment effect. Our results remain robust against different model specifications. Overall, the results underscore the broader societal benefits of home garden interventions in crisis settings, extending beyond direct nutritional improvements. The intervention demonstrated capacity-building and potential for enhancing household assets, contributing to livelihood recovery.
3 – Local Governance in Resiliency and Response to Poly-Crises in Mali
Jordan Kyle
International Food Policy Research Institute (IFPRI)
Abstract: In fragile contexts affected by simultaneous and overlapping environmental and social crises – such as droughts, floods, invasive pests, communal conflict, banditry, and extremist violence – local leaders play pivotal roles in shaping resiliency and crisis response. In a project in rural Mali conducted with the World Food Programme, we identify a variety of local leaders with different formal and informal arenas of authority that shape resiliency and response to poly-crises in Mali at the village and commune levels. Through an original survey of over 2,900 leaders across 556 villages in seven regions of Mali that face different constellations of poly-crises, we use experimental games, survey experiments, and survey modules on local crisis event histories to test three primary research questions. First, how do features of how anticipatory aid programming could be deployed – including the accuracy of early warning systems and the extent to which aid decision-making is devolved to the local level – influence local leaders’ receptivity to shifting resources from humanitarian response to anticipatory action programming? And, how do these effects vary with past exposure to shocks, exposure to different types of conflict, and variations in how effective local actors are at delivering aid? Second, how do local leaders view the costs and benefits of transparency over aid distribution, and how does this vary over the amount of discretion they have over its distribution? Third, how do each of these questions vary depending on whether leaders have weaker ties to current systems of aid distribution? Conducted in partnership with the World Food Programme, this research aims to inform a new aid strategy being developed in Mali to spend more on anticipatory action programming to prevent the worst outcomes from happening around flooding and droughts and to contribute to research on political economy of aid distribution.
To answer these questions, we implemented two experiments. The first used a cartoon storyboard to explain anticipatory action programming – aid that is triggered upon an early warning signal of a crisis to prevent the worst effects of a crisis – to leaders who may not be familiar with this type of programming. All leaders were provided with the same information on how anticipatory action (AA) works and some of the costs and benefits of allocating funds towards AA versus current models of humanitarian aid responding to crises after they have begun. We experimentally vary two aspects of the game: (1) the level of accuracy of early warning signals and (2) whether donors or local aid committees would be responsible for distributing aid after an early warning signal triggers the release of AA. After going through the storyboard, respondents play a budgeting game to indicate how much of 5 million CFA they would move from the “status quo” of aid response after crises over to AA. After making their allocation decisions, we ask respondents to close their eyes and draw a marble out of a jar, where the number of black marbles indicated how often the early warning signal was accurate out of 10 (remaining marbles were white) in the assigned treatment group. Afterwards, respondents are given the chance to reallocate their 5 million CFA between AA and humanitarian aid. The second, independent experiment is related to transparency and discretion over aid distribution. This experiment is designed as a discrete choice exercise where respondents are asked to select between two “aid profiles” three different times. Aid profiles randomly vary along three parameters: the degree of local discretion over targeting and distribution, the transparency provided over how households were selected and how logistics are handled, and the type of aid.