A ‘world’ model of integrated financial markets using artificial neural networks

Abstract

Over the last two decades, many important changes took place in the area of international finance. The liberalisation of financial markets and the development of powerful communication and trading facilities have enlarged the scope of selection for investors. Most financial markets of major developed countries must be regarded as highly integrated now. Traditional capital market theory has also changed and methods of financial analysis have been improved considerably. The existence of non-linearities in financial market movements has been emphasized by various researchers and financial analysts over the last years. Taking both aspects into account, a new kind of financial analysis seems to be necessary: the non-linear analysis of integrated financial markets. Recent developments in the theory of neural computation provide interesting mathematical tools for such a new kind of financial analysis. This paper presents both the economic approach to an analysis of highly integrated financial markets and the econometric methods, especially artificial neural networks (ANN), to realize it. Forecasting of any asset class, e.g. stock, bond or exchange rate prediction, now becomes an integrated part within such a wider global capital market equilibrium. ANN, especially recurrent networks, are used to model integrated financial markets. This approach tries to develop a ‘world’ model of integrated financial markets. The ‘world’ consists of the stock, bond and currency markets of the United States, Japan and Germany. Traditional econometric methods as well as ANN are used to develop various kinds of models, which in turn are rated by their ability to provide accurate forecasts for these financial markets in an out-of-sample test.