Estimating a Banking-Macro Model for Europe Using a Multi-Regime VAR

This paper introduces a Banking-Macro Model and estimates the linkages through a Multi-Regime VAR (MRVAR). We first introduce a dynamic model which is akin to the Brunnermeier and Sannikov (BS) model (2010). The banking sector borrows from capital markets, issues liabilities, accumulates assets, pays bonuses, but its equity value can undergo large fluctuations. The banking sector is exposed to an instability that not only arises from loan losses but also from adverse movements of asset prices and their impact on risk premia and credit spreads. We solve for the model numerically with different determinations of credit spread and different constraints on the banks`s decision variables. In contrast to the standard model of the financial accelerator, exhibiting mean reversion, our model, similarly to BS (2010), exhibits local instability. Whereas the standard model leads, in terms of econometrics, to a one-regime VAR we argue for the use of a MRVAR. We estimate our model for EU countries with a MRVAR using a constructed financial stress index and industrial production for those countries. We undertake impulse-response studies with a MRVAR which allows us to explore regime dependency of shocks. We show that the shocks have asymmetric effects, depending on the growth regime of the economy, and on the size of the shocks. Small financial stress shocks may not matter, but large shocks are likely to have magnifying effects. Stochastic dynamic modeling and Multi Regime VAR State and size dependency of positive and negative stress shocks