Housing-Rich, Cash-Poor: Balance-Sheet Structure and Recovery from Household Financial Fragility in China
Description
This paper examines household financial fragility as a dynamic process, focusing on entry, persistence, and recovery. Using four waves of the China Household Finance Survey (2015–2021), we define a household as financially fragile when its liquid assets are insufficient to cover three months of basic consumption. We first estimate adjacent-wave transition models and then apply a Wooldridge-style dynamic random-effects Probit to account for initial conditions and correlated household heterogeneity. Financial fragility is highly persistent but not absorbing: persistence exceeds entry by approximately 28–30 percentage points across transition periods, while roughly one-third to one-half of initially fragile households recover by the next survey wave. In the balanced-panel dynamic model, the average partial effect of lagged fragility falls from about 18 percentage points in a naïve random-effects specification to 5.9 percentage points after accounting for initial conditions and household heterogeneity, but remains statistically significant. Recovery also depends systematically on the depth of the liquidity shortfall and the composition of household balance sheets. Conditional on total assets and other household characteristics, a 10-percentage-point increase in the housing-asset share is associated with a 1.2-percentage-point lower probability of recovery, while housing-rich, cash-poor households have recovery probabilities about 8.6 percentage points lower than other fragile households. Higher leverage is likewise associated with weaker recovery. These findings show that observed persistence substantially reflects persistent household heterogeneity, while meaningful residual state dependence remains, and that balance-sheet wealth does not necessarily translate into liquidity resilience.