Risk Preferences Under Extreme Poverty : A Field Experiment

Until now, the dominant belief concerning the relationship between poverty and risk aversion is that the poor are more risk averse. If the poor are more risk averse, then they will choose “low risk-low return” activities that trap them in poverty. However, both empirical and experimental evidence show no clear pattern such as would suggest that the poor are somehow more averse to risk than others; at times, they even seem to embrace risk, while at other times, there seems to be no difference. Focus has tended to be on extreme behaviors, as these are related to sub-optimal decisions such as have even raised questions whether an individual can be simultaneously both poor and rational. Amongst all the available empirical evidence, there is one bit of evidence of special interest - changes in behavior whenever subsistence is at risk. This paper emerges from the fact that recent experimental evidence in both psychology and economics suggests that certain decisions made under risk respond to reference points. We develop a theory within the traditional stream of rational choices, whereby the references are set by only observable variables, such as prices and family size. According to this theory, extremely poor individuals respond to the income reference that guarantees the consumption of the necessary calories so as to ensure a healthy and longer life. Being in the neighborhood of this reference can incentivize both the seeking of high risk, whenever below the reference, and an aversion to high risk, when above. An experimental exercise was conducted involving 92 individuals from households living in poverty and extreme poverty wherein they participated in a baseline risk experiment that was the one we analyzed. Inasmuch as the treatment was not randomly assigned, but instead was determined based on households' per-capita incomes, a quasi-experimental approach was used to analyze the results. We use a regression discontinuity design, and find evidence suggesting that being presented with the opportunity of avoiding undernourishment significantly decreases a household's risk aversion