Abstract
The efficient market hypothesis (EMH) has long been perceived as the cornerstone of modern finance theory. However, the EMH has also recently been dismissed as "the most remarkable error in the history of economic theory" (Wall Street Journal, Oct. 23. 1987). Most of the early research was concerned with detecting the efficiencies or inefficiencies by autocorrelation tests, run tests, and filtering tests. In general, the inefficiencies detected are relatively small. Recently, however, there has been an explosion of research activity to detect inefficiencies in the general area which we call "nonlinear science". This dissertation aims at the applications of these kinds of new methodologies to the Swedish stock market and the Korean stock market. This dissertation consists of 8 chapters. Chapter 1 reviews the challenges to stock market efficiency, and chapter 2 criticizes traditional financial models and assumptions for the EMH tests. Chapter 3 discusses the sample data. In chapter 4, the estimated results under the product process model are presented. Chapter 5 is focused on the low power spectrum law. The power exponent is calculated for the samples. In chapter 6, the types of patterns (memory) are uncovered by Rescaled range (R/S) analysis. Chapter 7 deals with the market inefficiency arising from nonlinear dynamical systems theory. The BDS test for detecting nonlinear dependence is applied to the sample markets. Finally, chapter 8 summarizes the conclusions.