Eliciting risk preference in the field: Do methods and real incentives matter?

In economics, risk preferences are key elements of individual decision-making, and several methods are used to elicit them. We examine whether and to what extent risk elicitation methods and incentives matter when eliciting risk preferences in a lab-in-the-field experimental setting. First, considering results from three different incentivized risk elicitation methods, we compare preferences elicited using a single choice and those elicited from multiple decisions between structured pairs of lotteries. Second, using a multiple price listing (MPL) method with incentivized and hypothetical experimental settings, we examine if risk choice behavior in these two settings differs significantly or not. Third, we investigate (in)consistencies in risk preferences between experimental and survey measures. To investigate the above aspects, we conducted field experiments and surveys with 500 smallholder peri-urban farmers in Ethiopia. Findings suggest that different elicitation methods provide significantly different risk measures, though estimates are correlated across elicitation methods. Our results also indicate that there is a significant association between risk attitude behavior in incentivized and non-incentivized settings, implying that hypothetical experiments can also provide valid risk measures. Findings also reveal inconsistencies in risk preferences between experimental and survey measures, and such inconsistencies are partially explained by gender, interaction of gender and education, and trust. Opposite to our hypothesis, education alone does not explain inconsistencies between experimental and survey measures.