FORECASTING AUSTRIAN IPOS - AN APPLICATION OF LINEAR AND NEURAL-NETWORK ERROR-CORRECTION MODELS

In this paper we apply cointegration and Granger-causality analyses to construct linear and neural network error-correction models for an Au strian Initial Public Offerings IndeX (IPOX(ATK)). We use the signific ant relationship between the IPOX(ATK) and the Austrian Stock Market I ndex ATX to forecast the IPOX(ATX). For prediction purposes we apply a ugmented feedforward neural networks whose architecture is determined by Sequential Network Construction with the Schwartz Information Crite rion as an estimator for the prediction risk. Trading based on the for ecasts yields results superior to Buy and Hold or Moving Average tradi ng strategies in terms of mean-variance considerations.