سیاست های پولی مقاوم در یک مدل با بحران مالی

We characterise optimal discretionary monetary policy responses to cost-push shocks and to financial distress in the presence of model uncertainty. Under robust control, the central bank reacts more aggressively to both types of shocks, and less to the lagged policy rate, than if the true model is known. We document how the objective to stabilise the policy instrument conflicts with the concern for robustness to model misspecification: the higher the weight on interest rate stabilisation in the loss function, the more the robust policy deviates from the optimal policy under rational expectations. Financial distress is akin to a contractionary demand shock and does not induce a policy trade-off; thus model uncertainty does not constrain monetary policy in the face of financial shocks.

We characterise optimal monetary policy responses to cost-push shocks and to financial distress in the presence of model uncertainty. In particular, we examine the interaction between the policy maker’s concern for robustness and his desire to avoid excessive volatility in interest rates. We do so by including a penalty on interest rate changes in the central bank’s loss function, in addition to inflation and output gap volatility. There is no commitment device for future policy; the central bank sets the policy rate in a discretionary fashion. The analysis is based on Goodfriend and McCallum’s (2007) New Keynesian model with banking, where households need bank loans to make consumption purchases. Loan production uses labour and collateral as inputs. Financial distress is modelled as a negative shock to collateral. We adopt the robust control approach of Hansen and Sargent (2008). The true data generating process, unknown to the agents, lies in the neighbourhood of a so-called reference model. Facing Knightian uncertainty, the policy maker is unable to formulate a probability distribution over plausible models. A robust policy is one that performs well in the worst possible outcome of a pre-specified set of models. The paper aims to fill two gaps in the literature. First, robust monetary policy has not previously been applied to financial shocks. Second, the two policy objectives interest rate stability and robustness have not been analysed jointly in existing research. Our results can be summarised as follows. First, taking model uncertainty into account, the robust instrument rule implies stronger responses. The policy maker is more aggressive to both cost-push shocks and financial shocks. The more aggressive response is mirrored by a smaller degree of interest rate smoothing in the instrument rule. Second, the concern for instrument stability conflicts with the concern for robustness: the higher the weight on interest rate smoothing, the more the robust policy deviates from the optimal policy under rational expectations. Third, financial distress is akin to a demand shock in that it does not generate a policy trade-off. Consequently, the impulse responses to a financial shock in the worst possible outcome do not differ much from those under rational expectations. Thus, model uncertainty plays a minor role for policy in the face of financial distress. The remainder of the paper is organised as follows. Section 2 presents the New Keynesian model augmented with a banking sector. In Section 3, we analyse optimal robust monetary policy under discretion. As an extension, we consider in Section 4 an alternative policy objective including the external finance premium. Section 5 concludes.

This paper analyses optimal discretionary monetary policy under model uncertainty when the economy is hit by cost-push shocks and financial distress. Of particular interest is the interaction between a policy maker’s interest rate smoothing objective and his concern for robustness to model misspecification. The first objective is captured by a penalty on interest rate changes in the central bank’s loss function. To take into account model uncertainty, we follow the robust control approach, where the policy maker chooses a policy that performs well in the worst possible outcome. We compute the optimal implicit instrument rule under discretion. First, we find that the robust rule responds more aggressively to both cost-push shocks and financial shocks than the optimal policy under certainty equivalence. The increased aggressiveness is supported by a decline in the coefficient on the lagged policy rate. Second, the deviation of the robust policy from the rational expectations equilibrium widens for higher weights on interest rate smoothing in the central bank’s objective function. Thus, the policy maker’s desire to insure against model uncertainty conflicts with his concern for instrument stability. Third, financial distress is akin to a contractionary demand shock and as such does not induce a policy trade-off. As a consequence, the optimal dynamics following a financial shock are largely unaffected by the presence of model uncertainty.