LINEAR VERSUS NON-LINEAR PREDICTABILITY OF EQUITY PRICES: AN EMPIRICAL INVESTIGATION
- People & Global Business Association
- Global Business and Finance Review
- Vol.9 No.2
-
2004.1253 - 64 (12 pages)
- 8
The analysis of this study is twofold: 1) Using A non-linear test (the Hinich test), it provides evidence of non-linear dynamics in stock prices in the American market from January 1955 to December 2002. 2) Utilizing non-linear regression models, the Multiple Adaptive Regression Splines (MARS) and Ⅱ-Splines, the study examines the relationship between stock prices and their fundamentals. Non-linear results are compared with linear. The predictive ability of the non-linear models outperforms the linear. As a result, the study suggests that non-linear models should be considered for the analysis of stock prices.
Abstract
INTRODUCTION
METHODOLOGY, DATA ANALYSIS, VARIABLES IN THE SYSTEM, DEFINITIONS, HYPOTHESES TO BE INVESTIGATED
EMPIRICAL RESULTS
CONLUSION: THE ECONOMIC IMPORTANCE OF MARS MODEL'S FINDINGS
REFERENCES
(0)
(0)