This paper estimates the persistence of transient income shocks to farm households in rural Indonesia. Persistence is defined as the elasticity of a household's 1997 household per capita income with respect to its 1993 per capita income, controlling for time-invariant characteristics of the household. Local rainfall levels are used as an exogenous source of transitory variation in 1993 income. Four main conclusions emerge. First, roughly 30% of household income shocks remain after four years. Second, the persistence of negative and positive shocks is approximately equal; if anything, positive shocks last longer. Third, neither positive nor negative income shocks disproportionately affect poor households. Finally, measurement error in income and unobserved household heterogeneity are important sources of bias. These findings cast doubt on common arguments advocating public intervention to stabilize or redistribute income, and suggest that anti-poverty policy should address more permanent causes of household poverty. Copyright Blackwell Publishing Ltd 2005.