Background risk, insurance and investment behaviour: Experimental evidence from Kenya

We study in a lab experiment with slum dwellers in Nairobi whether background risk induces ex ante risk-avoiding behaviour. We find that risk-averse subjects choose indeed less risky, profitable investments when background risk is present. However, relaxing these risk constraints by the provision of insurance has no effect (free insurance) or even a negative impact (fair insurance) on risk taking. We find suggestive evidence that these unexpected results are driven by imperfect understanding of insurance and experience effects. Moreover, constantly low investment under fair insurance suggests a dampening effect of the certain premium that dominates the positive risk-reducing effect of insurance. Our results imply that providing poor households with instruments that facilitate the handling of background risk is important to encourage entrepreneurial risk taking. However, insurance might not be the optimal policy in the short term, since it may, if not sufficiently understood or valued, even reduce willingness to invest compared to the status quo.