Essays on threshold autoregressive modeling of bond time series.
Item
-
Title
-
Essays on threshold autoregressive modeling of bond time series.
-
Identifier
-
AAI3213146
-
identifier
-
3213146
-
Creator
-
Li, Jinghong.
-
Contributor
-
Adviser: Salih N. Neftci
-
Date
-
2006
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Economics, Finance
-
Abstract
-
The linear Gaussian models such as AR models, ARMA models and ARIMA models have been proved by many previous research that they are not ideally suited for modeling some financial time series that exhibiting asymmetry, limit cycles and jump phenomena, in other words, nonlinearity. Researchers therefore resort to nonlinear models to interpret financial time series. The threshold autoregressive model (TAR) introduced by Tong and Lim (1980), among the family of nonlinear models, has been found to best capture asymmetries, limit cycles and jump phenomena in the dynamic structure of economic and financial time series. Nonlinearity in stock prices and exchange rates has been often detected by various statistical tests. However, only few attempts have been made to subsequently model the nonlinearity explicitly. My dissertation investigates nonlinearity in bond yields and prices, and model the nonlinearity explicitly. The thesis consists of two essays.;In the first essay, I explore the presence of nonlinearities for the daily series of 10-year Japanese government bond (JGB) yields by using the Tsay (1989) test, I find the threshold nonlinearity due to a significant change in Japanese debt management policy. I test to find the sixth lag of the series as the threshold variable, then locate the threshold, and estimate a 2-regime self-exciting threshold autoregressive model (SETAR) for this time series.;In the second essay, I investigate the presence of nonlinearities for the daily series of 10-year American Treasury note (T-note) prices by using the Tsay (1989) test, I analyze the threshold nonlinearity due to government intervention and the price protection pursued by investors. I test to find the first lag of the series as the threshold variable, then I estimate two SETAR models for this time series based on different lag lengths and compare these two models.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
