Abstract

Estimating Sustainable Development Goals (SDG) indicators is essential to monitor progresses towards the UN 2030 Agenda on Sustainable Development. When census data is available, the complexity of producing such indicators rely on the availability of information and the calculating issues related to the definition of each of them. In several instances, this complexity is enhanced by the fact that data is collected on a sampling basis. In Latin America, approximately 1/3rd of countries conducts regular agricultural surveys, and among those all, all but two use a dual frame sampling design composed of one list frame, and one area frame. Throughout Europe and North America as well, multiple frame designs for agricultural surveys are the norm, and some African and Asian countries as well are adopting this approach. In multiple frame designs, the list frame component tends to cover large and/or specialized farms which do not change often, while the area frame component ensures adequate coverage of small holders avoiding the challenges of keeping an updated list. However, despite the many advantages in terms of coverage and ease to update, the use of multiple frames poses many challenges for estimating complex indicators such as SDG indicators 2.3.1 and 2.3.2 on small holder productivity and income, respectively. This paper directly addresses these challenges and develops strategies for accurately estimating SDG 2.3.1 and 2.3.2 using multiple frame sampling designs. Some practical solutions are provided to countries monitoring the SDGs, filling a gap in the methodological literature about these indicators.