Recent stress-simulations have been able to take into account the various, interacting contagion dynamics and network layers of financial systems, but can only classify networks' stability with regards to a specific initial stress-scenario. Approaches such as DebtRank provide a scenario-independent measure of systemic stability but only take institutions' solvency into account, even though liquidity and its interactions with solvency critically affect stability. We construct a liquidation-valuation shock-transition matrix that captures multiple, interacting contagion dynamics and network layers in a single, scenario independent measure of stability which allows us to study the pre-crisis build-up of financial distress. we show that even when financial systems have identical topologies throughout their layers, the various layer-specific contagion dynamics cause the systemic stability to depend simultaneously and heterogeneously on the various network layers. Furthermore, whereas leverage is regarded as a driver of instability, we show that the leverage in excess of the proportion of shock-absorbing institutions active in the funding and securities markets drives instability. We derive these insights from theoretical abstractions of financial systems, but calibration of the model to real data is straightforward, making it a valuable addition to the regulator's toolkit.