Essays on Communication and Behaviour Under Risk and Ambiguity

This dissertation consists of three chapters focusing on behaviour under risk and ambiguity. The first chapter analysed the best method to communicate risk information to weather forecast users whilst the last two analysed smallholder farmers’ and students’ decision making on crop selection when presented with uncertainty information of drought. In the first chapter, experimental economics methods were used to assess forecast user understanding of information in temperature forecast. We tested whether undergraduate students presented with uncertainty information (90th percent confidence intervals) in a table and bar graph format were able to correctly understand the forecast and use the extra information to choose the “correct" (most probable) outcome than if they are presented with a deterministic forecast. Participants from the University of Exeter were asked to choose the most probable temperature outcome between a set of “lotteries” based on the temperature up to five days ahead. If they chose a true statement, participants were rewarded with a cash payment. Results indicate that on average participants provided with uncertainty information performed better than those without. Statistical analysis indicates a possible learning effect as the experiment progressed. The second chapter assesses if there are gender differences in the behaviour of smallholder Zimbabwean farmers when faced with risk and ambiguity. The risk and ambiguity preferences of male and female farmers were elicited using a modified Holt and Laury (2002) field experiment. Farmers were asked to choose whether or not to adopt a new drought tolerant variety under different probabilities of a drought occurring. Subjects in one group were presented with known probabilities whilst another group was presented with ambiguous probabilities (range). Most of the farmers’ exhibited extreme ambiguity and risk aversion and female farmers were more averse. Results indicate heterogeneity and the need to disaggregate samples when analysing research results as there maybe underlying factors affecting different groups. The third chapter elicited the risk and ambiguity attitudes of vocational college students in Zimbabwe. Results indicate that in general, students were both risk averse and ambiguity averse. Those presented with the risk treatment were less risk averse compared with those shown the ambiguity treatment. Participants who were presented with the ambiguity treatment behaved as pessimists and perhaps made decisions based on probability of drought that was higher than the provided centre of the range. We found gender differences in risk attitudes: contrary to the norm, female participants were less risk averse compared to their male counterparts. This is however when all subjects are pooled together. Results also indicate that a higher certain payoff perhaps incentivises consistency and increases risk aversion. The data seems to indicate anchoring effects from varying the order the probability of drought was presented.