Doctor of Philosophy (PhD)



Document Type



This dissertation empirically identifies exogenous changes in monetary policy and estimates their effects on the economy. The framework is the Romer and Romer (2004) quasi-narrative approach to identifying exogenous changes in monetary policy. The first essay replicates the Romer-Romer (RR) “quasi-narrative” measure of shocks and updates them with Greenbook forecasts to 2003. A key result is the quasi-narrative approach is robust to updates and corrections for serial correlation. An alternative, independently formed measure of the intended funds rate from Thornton (2005) is compared to the RR measure. The measures are highly correlated and display slight differences concerning the timing of the changes in the intended funds rate. The second essay examines the relative importance of three sources of monetary policy shocks in the quasi-narrative approach. The sources analyzed are changes in operating regimes, changes in chairmen, and the credit controls of 1980. It is found that the responses of monetary policy to forecasted increases in output and inflation were strongest during the NBR targeting period and during the term of Paul Volcker as chairman. The most important source analyzed is shown to be the changes in chairmen. The third essay utilizes the quasi-narrative approach to create measures of monetary policy shocks from alternative real-time data. Three real-time data sources are constructed and explained. When jointly considered with the Greenbooks data used by RR, alternative real-time data is found to add significant information in the response of monetary policy. However, when compared to the RR results, the shocks produced from incorporating alternative data along with the Greenbook data produce only small and transitory differences in the responses of macroeconomic variables. Next, monetary policy shocks are constructed using only alternative real-time data that can be updated with a much shorter lag than is the case for shocks estimated using only Greenbook data. These new shocks are found to be highly correlated with the original RR measures. The shocks obtained from two specifications are shown to be reasonable substitutes for the RR measures, displaying only transitory, slight differences in the responses of output and prices.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

McMillin, William D

Included in

Economics Commons