Uncertainty, non-linear contagion and the credit quality channel: an application to the Spanish interbank market

Using granular data from the Spanish Central Credit Register, we study the contagion of financial distress via the credit quality channel in the Spanish interbank market. We propose a non-linear contagion mechanism dependent on banks’ balance-sheet structure (specifically, their leverage ratios). Moreover, we explicitly model uncertainty in lenders’ assessments of the probability of default of their borrowers, thus incorporating agents’ lack of complete information and heterogeneous expectations in their assessment of future outcomes. We perform multiple simulations across a wide range of possible levels of stress in the system, and we focus on disentangling the effects of these two key model components by comparing the results of our model with those of a linear and deterministic counterpart. We find that non-linear contagion leads to substantially larger losses than its linear counterpart for a wide range of intermediate levels of stress in the system, while its effects become negligible for very low and very high stress levels. Regarding uncertainty, we find that its effects, while smaller than those of non-linear contagion, are nonetheless relevant and most important around levels of stress at which different parts of the system become unstable. Interestingly, losses can be amplified or mitigated with respect to the deterministic case depending on the specific level of stress considered. Finally, the interaction between these two model components - non-linear contagion and uncertainty - alters the area where uncertainty matters.