Assessing the Vulnerability to Drought in Ceara, Northeast Brazil

Abstract

This paper seeks to contribute to the understanding of the current vulnerability to variability in rainfall of small farmers in the State of Ceará in Northeast Brazil by identifying factors that affect a household’s vulnerability and its ability to prepare for, or respond to, droughts. Understanding these factors can help to identify the households most in need of assistance during a drought and provide a target for programs that build capacity to adapt. Assistance programs currently use criteria to identify eligible programs. This analysis could also help inform an assessment of whether those criteria are appropriate for assisting the most vulnerable families. While, overall, we would expect that vulnerability would be tied to poverty, for this study, I make the assumption that within comparable levels of poverty, there are factors that may increase the resilience of some households to drought. Hence, although all households surveyed for this study would be considered poor, I assume that specific factors—such as land ownership, irrigation or pension income—may make a family more resilient to drought. Both qualitative and quantitative data and methods are used in this analysis. The qualitative section describes the information acquired in interviews with Brazilian experts. These responses were used to inform the selection of variables in the quantitative portion of the analysis. The quantitative piece uses data from a survey of small farmers in two areas of Ceará. Participation in a work front, a state-run temporary employment program available only during droughts, is used as a proxy for vulnerability to drought since historically work fronts have been used only as a last resort strategy when families have lost all other alternatives of income. A probit analysis is used to identify economic and demographic variables that are linked with increased probability of participation in the work front. A factor analysis is then used to identify components that explain a significant portion of the variance in the data set and to group households into similar groups. These results are then compared with the outcome of work front participation. Overall, the combination of probit and factor analyses proved successful in identifying households that were likely to have joined the work front in the past. The similarity of results from the probit and factor analysis supported the results. Additionally, variables identified as important determinants of vulnerability to droughts in interviews and the literature proved to be significant statistically. These key variables include on-farm production, non-farm income, particularly pensions, irrigation and plot size.