Many economic and financial time series data are available at high frequencies, as well as over long periods of time. Continuous-time regression models are clearly more suitable for such data sets. However, there still exists a paucity of methods developed to deal with continuous-time regression models. The first two chapters of my dissertation are devoted to develop two effective methodologies on continuous-time regression models to deal with high frequency data in the presence of non-stationary stochastic volatility and applies these methods to solve well known puzzles such as stock return predictability puzzle and the Uncovered Interest rate Parity (UIP) puzzle. As for the third chapter, we propose methods to forecast inflation. We demonstrate that using individual component information of the Consumer Price Index (CPI) helps to increase forecasting power of the CPI inflation rate.
The first chapter investigates a continuous-time parametric regression model in the presence of persistence and endogeneity of regressor and non-stationary stochastic volatility in error process. The technique consists of a simple time change to volatility time to accommodate a quite general form of stochastic volatility in error process and instrumental variable estimation to allow for a wide range of endogenous non-stationary covariate. We apply this method to test stock return predictability and cannot reject the null that stock returns are not predictable.
The second chapter studies a partially nonparametric regression model in continuous time. In this chapter, we propose a procedure to implement a semi-parametric estimation for a continuous-time regression model in the presence of non-stationary stochastic volatility. We adopt a series estimation method to model the non-parametric component in the regression. As for the non-stationary stochastic volatility, we again use a time change to transform the error process in our model to a Brownian motion. When we apply this method to empirical data, we cannot find any evidence to reject UIP for six major currencies vis- à-vis the US dollar.
The third chapter proposes methods for incorporating disaggregated components of the CPI to improve forecasts of the CPI inflation rate. We incorporate the CPI components directly in a dynamic regression model, as well as several methods of summarizing the information contained in the components. We find that the dynamic regression model using the Bayesian Information Criterion (BIC) to select the number of factors works very well. Overall, we conclude that CPI inflation forecasts can be improved by using summaries of information contained in the CPI components.
|Advisor:||Park, Joon Y.|
|Commitee:||Chang, Yoosoon, Escanciano, Juan Carlos, Walker, Todd B.|
|School Location:||United States -- Indiana|
|Source:||DAI-A 73/12(E), Dissertation Abstracts International|
|Keywords:||Cauchy estimator, Econometrics, Forecasting inflation, Martingale, Nonstationary stochastic volatility, Persistent regressor, Time change|
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