Essays on the coffee supply chain in Uganda

Doctor of PhilosophyDepartment of Agricultural EconomicsJason BergtoldCoffee is grown by many countries worldwide (Killeen & Harper, 2016). Africa contributes 12% to total world production (Wondemu, 2017). Uganda is the second largest producing country in Africa, after Ethiopia (Mwesigye & Nguyen, 2020). Coffee is a labor-intensive industry employing over 100 million people in 60 developing countries (Collinson et al., 2005). It has been particularly important to Uganda’s economy since 1961. It has employed over 12 million people (ICO, 2019) generating export revenues of US$ 492 million in 2018 (UCDA, 2018). The current annual value of export revenues from coffee exceeds US$17 billion (FAOSTAT, 2018), contributing to 1.5% of Uganda’s Gross Domestic Product (GDP) in 2019 (ICO, 2019). Uganda contributes over 31.24% of total coffee exports from Africa (FAOSTAT, 2018), placing it among the major coffee exporting countries in the world. Despite the paramount importance of coffee to Uganda’s economy, the industry still faces bottlenecks in the coffee supply chain that limit growth. The purpose of this dissertation is to examine and assess two primary bottlenecks in the Uganda’s coffee supply chain, productivity at the farm level and maintaining quality through the supply chain. The total area under coffee production in Uganda has increased over the past ten years with no substantial increase in productivity. Standard productivity measures that examine technical efficiency and productivity though are not able to \account for the age of coffee trees or plantations, which is a limiting factor in coffee productivity. Productivity assessments assume a homogeneous production frontier across farms, which may result in biased productivity estimates due to differences in age of trees on a plantation. The purpose of the first essay is to analyze coffee farm efficiency and productivity change over time accounting for the average age of trees on coffee plantations. Data for the study comes from the two main coffee producing countries in Africa: Uganda, and Ethiopia. Uganda is the primary study area, with Ethiopia being used for comparative purposes and to provide a foundation for generalization of research findings. World Bank data from the 2013/2014 and 2015/2016 growing seasons is used, comprising of 187 Ugandan and 606 Ethiopian farm households. Efficiency and productivity change are estimated using Data Envelopment Analysis (DEA) techniques to derive measures based on the Malmquist Index and its decomposition. This estimation is completed in two stages. In the first stage, technical efficiency scores are estimated separately using an unconditional DEA model and a conditional DEA model that accounts for differences in the average age of the coffee trees at the farm household level. The two indexes are compared to determine the impact of the age of the coffee tree on efficiency and productivity change. Finally, in the second stage, CMI scores are whitened and a nonparametric Kernel regression of land and labor on CMI is conducted to determine the impact of land and labor efficiency on productivity change. Estimating coffee farm productivity in the short run using unconditional scores results in biased productivity estimates and misleading conclusions. The average age of the coffee trees has a negative and statistically significant marginal effect on coffee farm productivity change in both Uganda and Ethiopia. As the trees get older, the efficient frontier retracts. This is important for development programs such as extension to identify the actual productivity loss due to managerial inefficiency. In addition, it provides evidence for the potential efficacy of coffee tree planting programs to help small famers. Increasing land input in coffee production decreases productivity due to thinning of other inputs, while increasing labor inputs improves productivity since this is a limiting factor in coffee production. The second essay analyzes coffee marketing channels and the current quality incentive structures to understand their impact on coffee quality through the coffee supply chain. This essay specifically considers two market channels through which exporters make transactions: the middleman market channel and the farmers’ group market channel. Data for this study comes from two main coffee producing districts in Uganda, Masaka and Mbale. Primary data were collected using pre-tested questionnaires. The data is comprised of interviews with 120 middlemen and 30 exporters, as well as four focus group discussions with producers. We use discrete choice methods based on actor decisions in the coffee supply chain to test for differences in quality of coffee transacted through the middlemen and farmers’ group channels. We apply the Principal-Agent framework to explain the impact of market channels on coffee bean quality. Results show a significant positive marginal effect of the market channel on coffee quality. If the market channel changes from middlemen to farmers’ group, the probability of the quality of coffee being high increases by 55 percentage points. The farmers’ group channel leads to high quality because of symmetric information between farmers, farmers’ groups, and exporters and availability of price incentives. The low quality through the middlemen channel is due to lack of incentives and information asymmetry between the farmers, middlemen and exporters. The government of Uganda can help improve coffee quality through promoting formation of farmers’ groups. This can be done through providing extension services to create awareness and infrastructural support to poorer and more remote farmers