Ebook Optimal Discretionary Policy vs Taylor Rule: Comparison under Zero Lower Bound and Financial Accelerator
In the modern monetary policy theory, interest rate policy rule is an important component of the model. Probably the most frequently used rule in the related literatures is the Taylor rule suggested by Taylor (1993) and its variations. Of course, this does not mean that the Taylor rule is the optimal policy rule. The optimal policy rule is obtained by solving the linear quadratic rational expectation problem in which the object function is the quadratic welfare loss function and the constraint is the New Keynesian Phillips curve(NKPC). The optimal policy is expressed as a relationship between output and inflation gaps and the interest rate is determined with intertemporal IS curve. Optimal policy is divided into optimal discretionary policy and optimal commitment policy depending on whether the central bank can commit on their policy in the future. Optimal discretionary policy is time consistent by construction and optimal commitment policy is time consistent by assuming commitment technology imposed on the central bank. In this paper we will examine only the optimal discretionary policy because of computational burden in case of the optimal commitment policy.
Then, which interest rate rule is followed by the central bank in reality? Is optimal policy rule always desirable than Taylor rule? To answer these questions, we extend the traditional dynamic new Keynesian(DNK) model by introducing the zero lower bound(ZLB) constraint and financial accelerator. The reason for extending the model in these directions is twofold.
First, with this extended model, we expect the different effects of the rules on the economy will be revealed distinctively. A peculiar fact in the literatures of ZLB constraint is that they focus on demand shock. This can be understood by noting that the policy problem of liquidity trap occurs only with demand shock. When the ZLB constraint is not considered, the interest rate policy implied by optimal discretionary policy in response to demand shock is trivial and it is not different from the policy implied by the Taylor rule qualitatively. Both of policy rules dictate to raise (lower) the interest rate when there is a positive (negative) demand shock. However, the ZLB constraint changes the shape of optimal discretionary policy significantly, as is shown in figure 1. It is because the ZLB constraint makes the optimization problem to be nonlinear. The solution to this nonlinear problem with respect to demand shock is no longer trivial and makes the interest rate policy quite different from the Taylor rule. In addition, the existence of financial accelerator causes the qualitative difference between the two policy rules.
Second, the ZLB constraint and financial accelerator have substantive meaning related with the recent global economic crisis. The crisis started from the financial sector and many central banks around the world responded to the shock by dropping the short term policy interest rates precipitously near to the zero rate. With traditional DNK model, these substantive issues cannot be dealt with. Representative agent and frictionless financial transaction assumptions fundamentally prevent the consideration of financial sector problem in the model. The possibility of ZLB constraint is also ignored as the model is constructed by linearizing the first order optimization conditions at around the steady state. Hence, we expect the extended model to provide more substantive analysis environments.
The problem of ZLB constraint, or liquidity trap, was studied in many literatures at around 2000: Krugman (1998), McCallum (2000), Eggertsson and Woodford (2003), Svensson (2000), to name a few. The main interest of those papers was how to escape from the liquidity trap with monetary policy, which seems to be impotent with the interest rate reached on the zero bound. Without considering the role of financial sector in the model,5 the solution focused on how to eliminate the deflation expectation. They emphasized the commitment of the monetary policy to the inflation in the future. Recently, Jung, Teranishi, and Watanabe (2005), Nakov (2008), Adam and Billi (2004), and Adam and Billi (2006) tried to assess the optimal monetary policy under zero lower bound constraint. Noting that the optimal monetary policy problem with ZLB constraint is nonlinear problem, they showed that the solution of the problem is quite different from that of the problem without ZLB constraint as the public recognize the limited ability of the central bank when the economy reaches the zero bound. The equilibrium of the economy is different depending on the possibility of the policy commitment not only in the analytical solution forms but also in the observational results.
Contents
1 Introduction
2 Literature Review
3 Model
4 Computational Method and Calibration
5 Computation Results
- 5.1 Effects of Financial Accelerator
5.2 Effects of Financial Shock
5.3 Welfare Effect of Policy Regimes
5.4 Taylor Rule Augmented with Risk Premium
6 Conclusion
A Sensitivity tests
B Parameter estimation
Posted in :
