Abstract
Climate change is one of the most pressing issues of our time; it threatens the lives and livelihoods of many people around the world. The effects of climate change are global in scope and unparalleled in scale, ranging from extreme weather that disrupts harvests to rising sea levels that raise the risk of catastrophic flooding. Adapting to these effects in the future will be more difficult and expensive if significant actions are not taken today. In Africa, the impact of climate change is growing; it affects the most vulnerable the hardest, and contributes to food insecurity, displacement of populations and scarce water resources. Identifying programmes at global level and even at country level that address mitigation and adaptation to climate change is important to achieving our climate goals. In line with this concern, this thesis investigates the effect of weather variability on agricultural output in East Africa. The thesis attempts to answer the questions of whether there is a specific threshold beyond which variability in seasonal temperature and precipitation becomes harmful to agricultural production. If it exists, does agricultural production co-move with extreme weather beyond that specific threshold? Does variability in temperature and precipitation have a short- or a long-run effect? The answers to these questions could help us improve food production in East Africa.
We found a significant negative effect on agricultural output when rainfall variability exceeds the threshold of -0.533 ml and -0.902 ml in spring and summer respectively. Once the variability in temperature and precipitation exceeds the threshold, the dependence of yield and production of crops becomes stronger than in normal conditions; an increase in temperature or a decrease in rainfall is associated with a decrease in crop yield or production. We found that variability in temperature has a long-run effect on agricultural output, while variability in precipitation has a short-run impact. The thesis comprises three main chapters. Agricultural data were collected from the Food and Agriculture Organization (FAO) and climatic information from the Climate Research Unit (CRU) from 1961 to 2016. The first empirical chapter uses the panel threshold model to investigate the threshold at which variability in weather variables is harmful to agricultural output in East Africa. The findings were that growing rainfall variability has significant effects on agricultural output. More specifically, we found that rainfall variability in spring and summer is harmful to agricultural output when precipitation variability exceeds thresholds of -0.533ml and -0.902ml respectively. This implies that not all variability observed in rainfall has a negative effect. We found no significant
IV
effect in fall. In the case of growing-season temperature variability, we found no significant effects.
The second empirical chapter investigates the effect of extreme weather variability on the agricultural crop-related variables yield and production, over different periods of time. Events such as heat-waves and droughts are examples of phenomena that are more unusual and severe than regular or ordinary weather. According to the Intergovernmental Panel on Climate Change (IPCC, 2007; 2012) and Niang et al. (2014), the intensity and frequency of extreme events are both predicted to rise in East Africa. To capture when weather variability moves into the arena of extreme-weather events, and their impact on crop yield and production, a Markov-switching time-varying copula was chosen. Climate data are divided into higher and lower regimes. The abnormal or higher regime is the period during which the temperature not only exceeds a certain threshold. But moves into the category of ‘extreme event’, while the opposite holds for precipitation in the lower regime. The findings were that there is a strong dependence between weather variability and crops, meaning an increase in the temperature or decrease in rainfall is associated with a decrease in crop yield or production. The dependence is more significant when weather variability moves into either of these regimes compared to the normal condition. Given that the second chapter found there is high dependence between agricultural products and weather variability, there is a necessity to investigate whether this effect occurs in the short or the long run. For instance, an important distinction to notice is the difference between short-term and long-term adaptations. Capital and resources are generally fixed in the short term, and most potential modifications are restricted to behavioural changes. Over time, all capital resources are replaced or modified. This implies that adaptation is more feasible in the long run than in the short run. This last chapter assesses whether the effects of variability in temperature and precipitation on agricultural output are short- or long-run effects. If a cross-sectional dependence exists between the cross-sectional countries under investigation, the first generation of panel data techniques is not applicable. We employed the pooled mean group model and dynamic fixed-effect model, and found that variability in temperature has a long-run impact on agricultural output, while variability in precipitation has a short-run effect. In addition, the effect is heterogeneous between countries.
Keywords: weather variability, agricultural output, East Africa.