`

Natural Shocks and Farm Household Behavior:
the Consequences of an Earthquake in PISO Peru


Marrit van den Berg, Ricardo Fort

Wageningen University, Wageningen, The Netherlands
GRADE, Lima, Peru



Abstract:
The livelihoods of rural households in developing countries are constantly affected by their exposure to natural hazards. Shocks like earthquakes may destroy livelihoods and people in risk-prone areas will engage in a number of both ex ante and ex post coping strategies. These strategies are costly, especially so for poor people with limited access to financial and insurance markets. While wealthy households may be able to cope with fluctuations in income through (dis)savings, poor households in volatile environments will need to adopt a defensive portfolio strategy to smooth income and assets. They will diversify or skew towards low-risk low-return alternatives that avoid destitution but at the same time inhibit income growth and investment. Individual shocks may strengthen this tendency. A better understanding of household’s attitudes towards risk, as well as ex-ante actions taken to prevent losses and ex-post coping strategies can provide governments and donor’s institutions with better policies and instruments to reduce household vulnerability and increase income growth. We use a unique data set covering survey and experimental data of cotton-farming households before and after the 7.9 degrees in the Richter scale hit that the Valley of Pisco and surrounding areas in 2007. The 2006-2009 panel covers 72 affected households and 107 controls, which were selected from a larger sample of the base line survey using propensity score matching. We can therefore not only analyze changes in the households affected by the earthquake but also compare these changes in comparable households that were not affected. The questionnaires include sections on household composition, income generation activities, and the willingness to purchase insurance. Moreover, we performed hypothetical experiments to measure risk attitudes. The paper will test the hypothesis that the Pisco earthquake has made people more careful and more willing to take (costly) risk management measures, such as the purchase of insurance. A shock like an earthquake may not only destroy assets and result in income losses but also affect household perceptions and/or preferences: the experience of a disaster may make people more careful and inclined to engage in costly risk-coping or management strategies. We compare changes in choices in a simple risk experiment and preferences for insurance as well as actual portfolio choices, such as the choice of cotton versus maize and diversification into off-farm activities. Preliminary analysis indeed shows that the earthquake has negatively affected incomes and has increased risk-averting behavior.

Key words: risk, experiments, panel data, Peru