PUB - Publikationen an der Universität Bielefeld

Recent papers point to the problem that inflation-targeting models do not as of yet consider financial market stability that can considerably derail inflation-targeting monetary policy, implying significant nonzero crisis probabilities that could come along with large negative output and employment gaps. Credit flows and the instability of credit appear to be at the root of the financial instability problem. On the other hand, some authors recently questioned whether a too early and too strong leaning against the wind policy by central banks might have higher costs than benefits in terms of output and employment losses. In our paper, we include in an inflation targeting model a financial stabilization goal. In contrast to infinite horizon and two-period models, we propose a finite horizon model. The model is solved by using a new global solution algorithm, called Nonlinear Model Predictive Control (NMPC), exploring stabilizing/destabilizing effects of price and nonprice (credit volume) drivers of the output gap, inflation, and credit flows. We substantiate the theoretical part of the paper by approaching the subject empirically, relying to that end on a regime-switching structural vector autoregressive (VAR) for the euro area. The empirical model contains standard macroeconomic variables along with credit flows and loan interest rates, the central bank policy rate, and European Central Bank (ECB) balance sheet variables. The regime-switching feature of the model is meant to capture the state-dependent relationship between the variables, with specific nonlinearities having direct counterparts in the theoretical model. Based on a sign restriction methodology, we explore conventional and unconventional monetary policy shocks, loan supply, and demand shocks, under different regime assumptions to reveal the state-dependent effects of both interest rate and volume-based policies. The empirical results are used as guidance for the calibration of the theoretical model variants.