This dissertation presents two essays and explores macroeconomic shocks' effect on the U.S. monetary policy regime change and company news' impact on the U.S. stock returns. They both concern the idea of agents' beliefs in forming their expectations about monetary policy behavior and stock market movements. Both pieces extensively apply econometric models in testing the underlying economic mechanisms with roots in theories.
The first essay studies the fundamental causes of the monetary policy regime switches within rational expectations models. I introduce a threshold-switching monetary policy process into the model that links the policy stance to the fundamental shocks by an autoregressive regime strength index. It creates an expectations feedback mechanism between private agents' policy forecasts and future policy regime outcomes. As demonstrated in a novel threshold-switching Fisherian model, well management of the private sector's expectations of policy regime change can have the same effect as actually switching the regime. Contrastingly, failure to do so leads to unfavorable outcomes of policy intention. Then, I embed the new mechanism into a New Keynesian model. Along the way, I also develop an efficient non-simulation based threshold-switching Kalman filter, in conjunction with a solution method that accounts for the endogeneity of switching regimes, to estimate the nonlinear New Keynesian model. My key empirical findings are threefold. First, non-policy shocks have been instrumental in driving U.S. monetary policy regime changes during the post-World War II period. Most notably, markup shock explains 65.6% of variations in the policy regimes. Second, absent from markup shocks, eight of the eleven less aggressive regimes would not have happened during this history. Finally, I conclude that linking the private sector's dynamic expectations formation and the Fed's dilemma of the dual mandate in the presence of adverse supply shocks is a promising path towards providing micro-foundations for monetary policy regime shifts.
The second essay systematically investigates how firm-specific news articles, obtained from Thomson Reuters, affect the future abnormal returns of U.S. stocks. Utilizing machine learning models, I find that adding information from news articles significantly improves the out-of-sample prediction performance, compared with using only stock prices and trading volumes as predictors. The improvement is economically and statistically significant across forecasting horizons from a month to half a year. Several information transmission mechanisms contribute to the news' predictive power. Media bias, measured by news sentiment, contains fundamental information about firms' values that are not yet embedded in stock prices. Specifically, it affects stock returns through the sensationalism of news and disagreement among media. Investor psychological bias also adds to market inefficiency. Investors seem to be overconfident, and it causes an initial overreaction to the news events. Investors also tend to purchase stocks with broader media visibility blindly and overreact to old news. Machine learning models excel at extracting the nonlinear interactions among news features and generate an increase of 0.23 in information ratio against S&P 500 index during the sample period.
|Advisor:||Glomm, Gerhard, Xu, Ke-Li|
|Commitee:||Walker, Todd B., Schneemeier, Jan|
|School Location:||United States -- Indiana|
|Source:||DAI-A 82/1(E), Dissertation Abstracts International|
|Keywords:||Cross sectional stock returns, Expectations formation effects, Information diffusion, Machine learning forecasting, Monetary policy, Regime switching|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be