A Threshold Model of Monetary Policy
Michael D. Bradley, Dennis W. Jansen
Abstract
Michael D. Bradley, Dennis W. Jansen
Abstract
Identification of the monetary policy process is of ongoing interest to both economists and private sector agents. Econometricians have used a variety of tools to identify which macroeconomic variables stimulate a change in monetary policy and to estimate the magnitude and timing of the response. The overwhelming majority of these attempts are both linear and symmetric, assuming that Federal Reserve response to macroeconomic conditions is the same in expansions in contractions. In addition, most of the literature assumes a simple and unchanging dynamic response in the policy variable to changes in the economy. Consequently it has been difficult for econometricians to find a structural monetary policy equation that accurately represents policy responses over wide range of economic conditions. The result is application of a variety of dummy variable and period-splitting analyses in an attempt to control for variations in the policy response. In this paper, we use a smooth-transition threshold model to extend this literature. This model identifies the threshold values for key variables, like inflation and unemployment, at which monetary policy becomes more active and the estimates both the active period dynamics and the transition process into and out of that period. This latter function provides a direct estimation of the degree of interest rate smoothing without the need for assuming and ad hoc “concern†for smooth rates. For example, the threshold model accounts for a moderate Federal Reserve response to changes in inflation in low inflation periods and a more rapid and larger response in high inflation periods. A similar analysis can be done for the unemployment rate. Finally, we test to see if monetary policy operates under different dynamics during periods of “stagflation†in which both the inflation rate and the unemployment rate is high. Finally, we estimate the degree to which monetary policy response is asymmetric. We explicitly estimate whether increases and decreases in key variables like inflation and unemployment lead to monetary policy responses of equal magnitude and timing.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Identification of the monetary policy process is of ongoing interest to both economists and private sector agents. Econometricians have used a variety of tools to identify which macroeconomic variables stimulate a change in monetary policy and to estimate the magnitude and timing of the response. The overwhelming majority of these attempts are both linear and symmetric, assuming that Federal Reserve response to macroeconomic conditions is the same in expansions in contractions. In addition, most of the literature assumes a simple and unchanging dynamic response in the policy variable to changes in the economy. Consequently it has been difficult for econometricians to find a structural monetary policy equation that accurately represents policy responses over wide range of economic conditions. The result is application of a variety of dummy variable and period-splitting analyses in an attempt to control for variations in the policy response. In this paper, we use a smooth-transition threshold model to extend this literature. This model identifies the threshold values for key variables, like inflation and unemployment, at which monetary policy becomes more active and the estimates both the active period dynamics and the transition process into and out of that period. This latter function provides a direct estimation of the degree of interest rate smoothing without the need for assuming and ad hoc “concern†for smooth rates. For example, the threshold model accounts for a moderate Federal Reserve response to changes in inflation in low inflation periods and a more rapid and larger response in high inflation periods. A similar analysis can be done for the unemployment rate. Finally, we test to see if monetary policy operates under different dynamics during periods of “stagflation†in which both the inflation rate and the unemployment rate is high. Finally, we estimate the degree to which monetary policy response is asymmetric. We explicitly estimate whether increases and decreases in key variables like inflation and unemployment lead to monetary policy responses of equal magnitude and timing.
Key concepts: Monetary policy, Economics, Inflation (cosmology), Interest rate, Variable (mathematics), Econometrics, Unemployment, Inflation targeting