Taylor-Based Monetary Policy Rules: Are They Forward-Looking, Data Congruent, and Asset Price Responsive?
Document Type
Doctoral Dissertation
Degree Name
Doctor of Philosophy
Publication Date
2003
Granting Institution
Western Michigan University
Abstract
This dissertation presents three essays to analyze a class of Taylor-based monetary policy rules that forms the basis of contemporary monetary policy decisions. The first essay examines whether policy rules are forward-looking. It proposes the methodology that, if the parameters of the Taylor rule change when the mechanism generating inflation changes, that is the Lucas critique applies, then inflation is not superexogenous for the parameters of the rule. In this case where superexogeneity fails, the rule is forward-looking. However, the results indicate that, although there is structural break in the inflation volatility process as evidenced by a model of heteroskedastic variance that allows for discrete regime shifts, we fail to reject the null of superexogeneity. This implies that there is insufficient evidence to reject the claim that Taylor rules are not forward-looking.
Because of the lack of power of superexogeneity tests, the second essay uses a more direct approach to analyze the encompassing properties of Taylor rules. It focuses on a set of non-nested interest rate rules to identify which model appropriately characterizes the interest rate data generating process. Five inflation and four output gap measures are studied and 116 non-nested P tests are conducted to show that the Philadelphia Fed expected inflation and the Congressional Budget Office output gap are the measures that are most consistent with the Fed Funds rate behavior. The model based on these indicators alone encompasses competing alternatives and is data congruent.
The third essay augments the model from the second essay to investigate whether the Fed reacts to the stock market while designing monetary policy. It employs a non-linear model of intrinsic bubbles to isolate market fundamentals from stock prices, and an Exponential Generalized Autoregressive Conditional Heteroskedastic model to gauge volatility clustering and leverage effects inherent in equity returns. These constructed measures of asset bubbles, fundamental values and volatility are then used to show that the Fed lowers interest rates in response to an increase in volatility and uncertainty, and an improvement in market fundamentals. Moreover, it accommodates productivity gains. However, there is no evidence that the Fed responds to asset price bubbles.
Rights
Copyright 2003 by Sooreea, Rajeev All rights reserved.