COVID-19 lockdown duration and transitions in family intergenerational support patterns: Evidence from the China Family Panel Studies

The family is the primary informal insurance mechanism through which economies buffer individual and aggregate risks, yet whether a major crisis activates intergenerational mutual aid or, instead, weakens intergenerational ties through physical isolation and resource strain remains unclear. Existing studies mostly examine changes in a single support dimension in isolation, or the simple before-versus-after correlation, and often rely on small, short-run samples, leaving the overall effect of the shock on the configuration of intergenerational support poorly understood. Using four waves (2016–2022) of the China Family Panel Studies (CFPS), we first classify intergenerational support into four cross-wave comparable patterns—Reciprocal, Filial-support, Interactive, and Detached—through latent class analysis, and then apply a multi-group difference-in-differences design to identify the causal effect of lockdown intensity on transitions between support patterns, its dynamic path, and its heterogeneity by household resources. We find that prolonged lockdowns shift households away from one-way support or detachment toward two-way reciprocity: the effect increases in lockdown intensity, is concentrated in the most strictly locked-down regions, and raises the post-pandemic probability of being Reciprocal by about 4.8 and 5.1 percentage points on average. The effect is delayed, being close to zero in the shock year (2020) and significant only by 2022. The shift toward reciprocity is widespread across income groups, whereas transitions into Filial-support are highly concentrated among low-income households. Overall, a major public-health shock triggers an adaptive reconfiguration of intergenerational support patterns, but its direction, lag, and form depend on shock intensity and household resources; the family, as an informal safety net, is resource-asymmetric, all the more so for vulnerable households.