Abstract
Climate change poses a serious threat to the growth of developing countries, especially Ghana where the majority of people derive their livelihoods directly from agriculture and related industries. This thesis sets out to examine the economic impact of climate change on agriculture by modelling climate, household, and farm data from Ghana. The impact is assessed through three connected empirical studies.
In our first empirical study, we estimate a multinomial logit (8,700 observations) in order to determine the factors that influence the choice of farming systems in Ghana. Consistent with our expectations, we find that climate is an important determinant of farm selection. Dry conditions (i.e. an increase in temperature or a decline in rainfall) favour the selection of livestock farms and mixed farms (i.e. mixed food-crop and livestock). Wet or cool conditions (i.e. a decrease in temperature or an increase in rainfall) favour the selection of tree-based farms. A decrease in temperature also favours the selection of food-crop farms. Based on the multinomial estimates and various projections of future climate, we simulate the potential impact of climate change and find that farmers will likely adapt by switching from tree-based farms (a highly profitable farming system) to less profitable but climate-resilient farming systems such as livestock farms.
In our second study, we use a flexible structural Ricardian model, SRM (which is a simultaneous two-stage optimisation technique) to estimate the impact of climate change on food-crop production by relying on 6,400 observations. In our version of the SRM, we control for temperature-rainfall interaction in the first stage and then estimate the second stage semi-parametrically. We find that rainfall impacts positively on the productivity of all food-crops except millet. Temperature has a negative effect on the productivity of most food-crops. A simulation of the effects of climate change shows that crop farmers will likely adapt by replacing high-value crops with millet, a low-value but climate-resilient crop. All things being equal, the results of our first two studies imply a decline in the aggregate value of agricultural output, hence there is the need to invest in research that seeks to improve the climate resilience of high-value food-crops and tree-based farms.
In our final study, we apply a Heckman selection model to 10,200 observations in order to fit Ghanaian farm and non-farm incomes and on that basis, simultaneously estimate the impact of temperature on farm income, non-farm income, and real food consumption. As expected, we find that income determines real food consumption. We find an inverse relationship between farm income and non-farm income. Warming impacts negatively on both farm and non-farm productivity and consequently real food consumption. For a typical adult, a 1oC rise in temperature results in a 4% reduction in real food consumption. This result has important implications for food security and general welfare.