Saving response to unemployment of a sibling

• This paper tests if a household's saving increases after a sibling's unemployment.
• Empirical results support this learning hypothesis.
• Such learning is estimated to raise the saving rate by 2–3 percentage points.
• Simulations also confirm that such learning can raise precautionary saving.

Standard theoretical models of household saving behavior do not typically assume that household perceptions of the world change in response to observed events. In light of the potential importance of such perception changes (e.g., after a financial crisis), this paper considers the hypothesis that a household's saving rate rises through informal learning after a sibling (direct or in-law) has been unemployed. The empirical results in this paper are consistent with the learning hypothesis, with coefficients estimated by the instrumental variable (IV) method implying that a household's saving rate increases by 2–3 percentage points if a sibling has been unemployed.