Continental lithospheric mantle (CLM) may have been built from subducted slabs, but the apparent lack of concurrent oceanic crust in CLM, known as the mass imbalance problem, remains unresolved. Here, we present a simple dynamic model to evaluate the likelihood of losing dense eclogitized oceanic crust from CLM by gravitational instability. Our model allowed us to assess the long-term evolution of such crust removal, based on how thermal and viscosity profiles change over time across the continental lithosphere. We found that the oceanic crust incorporated early into CLM can quickly escape to the asthenosphere, whereas that incorporated after a certain age would be preserved in CLM. This study provides a plausible explanation for the mass imbalance problem posed by the oceanic ridge origin hypothesis of CLM and points to the significance of preservation bias inherent to the studies of cratonic diamonds.