Using a unique household-level dataset on the situation after the Kobe earthquake in 1995, we test the full consumption risk sharing hypothesis, relaxing the separability assumption, and examine households' simultaneous choice of risk coping measures. Using multivariate probit estimations, we find that the full consumption insurance hypothesis is strongly rejected and our results indicate that households' utility across different expenditure items is not separable. As for households' choice of risk-coping measures, households borrowed extensively against housing damage, but relied on dissaving to cope with smaller asset damage, implying a hierarchy of risk-coping measures from dissaving to borrowing.