Files in this item
|(no description provided)|
|Title:||Experience in the application of unit roots and fractional difference models and tests|
|Author(s):||Agiakloglou, Christos N.|
|Doctoral Committee Chair(s):||Newbold, Paul|
|Department / Program:||Economics|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||One of the most important aspects in analyzing economic time series is to specify whether the observed series is generated by a stationary or non-stationary process, since most macroeconomic variables could be generated by a unit autoregressive root process. This determination as to whether or not a series should be differenced is known as the unit root test. In other words, if a time series in non-stationary, then in the presence of a unit autoregressive root, it can be converted to a stationary and invertible process by first differencing.
This thesis examines some techniques for testing the unit root hypothesis. It evaluates first the Dickey-Fuller-Type Tests for a unit autoregressive root in the presence of moving average components for any ARIMA (p, 1, q) process, proposed by Said and Dickey (1984, 1985). This type of test is conducted by either fitting a long autoregression to the data when the orders p and q are unknown, or by using the one-step Gauss-Newton least squares estimation when the orders p and q are known. Furthermore, to investigate the performance of the above tests, some applications of unit root models presented in the recent literature are also re-examined.
Moreover, this thesis employs the analysis of fractionally integrated models, which nests the unit root phenomenon, to distinguish in a more general way the value of the difference parameter. First, it evaluates and provides examples of the Geweke and Porter-Hudak (1983) estimation procedure which is frequently used to obtain an estimate of the fractional difference parameter, d. Next, it presents some new alternative testing techniques for fractionally integrated models. Finally, the behavior of the sample autocorrelations of fractional white noise is investigated.
|Rights Information:||Copyright 1992 Agiakloglou, Christos N.|
|Date Available in IDEALS:||2011-05-07|
|Identifier in Online Catalog:||AAI9305446|