DO METHODS MATTER? IMPLICATIONS FOR UNDERSTANDING AGRICULTURAL PRODUCTIVITY DRIVERS IN SMALLHOLDER FARMING SYSTEMS

The effects of the structure and efficiency of agricultural systems are pervasive throughout developing economies, reaching numerous sectors including food security, health, economic growth, income inequality, poverty and vulnerability, and gender equality, among others. Despite the ubiquitous impacts of agricultural productivity on various indicators of wellbeing, realized productivity is generally well below its potential. Increasing agricultural productivity, therefore, is central to many poverty reduction policies. Policy aimed at increasing agricultural productivity is often informed by household surveys that rely on self-reported information. Measurement error in household survey data could potentially lead to inappropriate or insufficient understandings of the drivers of agricultural production and the policy levers that could be used to improve productivity.I hypothesize that measurement error in household surveys, including that resulting from inadequate survey methodology, alters our understanding of certain agricultural productivity relationships. Unique survey data from Uganda and Ethiopia, which included the measurement of key agricultural inputs and production through both traditional self-reported methods as well as utilizing highly advanced data collection methodology, allows for unprecedented agricultural analysis. This dissertation, which takes the form of three essays, examines the following questions: (i) can the commonly observed inverse farm size-productivity relationship be explained by measurement error in production measurement?; (ii) how is agricultural productivity hindered by crop-specific soil suitability, and how are productivity constraints distributed across the agricultural population?; and (iii) does high-resolution plot-level soil data reveal gender-differentiated land quality endowments, and does this explain observed gender-based productivity gaps? Through this research I illustrate that methods do matter. The controversial inverse farm size-productivity relationship vanishes when objective methods of measuring crop production and plot area are used in place of farmer estimated production. Analysis of crop-specific soil suitability reveals additional nuances of agricultural productivity that are otherwise undetectable with farmer-assessment of soils, highlighting limited potential production gains for maize farmers in Eastern Uganda. Finally, utilization of high-resolution plot-level soil data vis-à-vis geospatial soil data uncovers gender-based differences in soil quality endowments, which, when included in productivity analysis, are shown to make up as much as 30 percent of the total gender-based productivity gap in Malawi.