The Consumption Insurance Paradox

The paper presented in this Manneim Appled Seminar studies consumption insurance heterogeneity across the income distribution in China. Standard theory predicts that wealthier households smooth consumption more effectively, given their lower income volatility and greater access to financial markets. Using seven waves of the China Family Panel Studies (2010–2022) and the Blundell, Pistaferri and Preston (2008) estimator, the author finds the opposite result. Poor households self-insure a substantially larger fraction of permanent income shocks than wealthy households, with a gap that is both statistically significant and large relative to cross-country benchmarks. A fixed-effects regression confirms the pattern using a different identification strategy. The paradox is concentrated at the permanent shock margin; differences in transitory shock insurance across groups are small and insignificant. He examines three candidate explanations. Consumption rigidity near a subsistence floor is ruled out: poor households' consumption responds symmetrically to income increases and decreases. Differential formal insurance access is ruled out, as the poor-wealthy gap is nearly identical in both urban and rural subsamples. Income composition provides a partial explanation: wealthy households hold more volatile non-wage income, which inflates their estimated permanent shock pass-through. The findings have two policy implications. First, expanding formal insurance programs may not close consumption smoothing gaps across income groups, since the paradox persists even where formal coverage is high and similar across groups. Second, targeting social insurance based solely on income may overlook important heterogeneity in actual insurance needs, particularly if the measured pass-through does not align neatly with smoothing capacity.