Forecasting stock market averages to enhance profitable trading strategies

Arbeitspapier

In this paper we design a simple trading strategy to exploit the hypothesized distinct informational content of the arithmetic and geometric mean. The rejection of cointegration between the two stock market indicators supports this conjecture. The profits generated by this cheaply replicable trading scheme cannot be expected to persist. Therefore we forecast the averages using autoregressive linear and neural network models to gain a competitive advantage relative to other investors. Refining the trading scheme using the forecasts further increases the mean return as compared to a buy and hold strategy.

Object Details

• Language: Englisch

References and Relationships

• Bibliographic citation: Series: Reihe Ökonomie / Economics Series ; No. 21

Classification and Topics

• Classification: Wirtschaft ; Information and Market Efficiency; Event Studies; Insider Trading ; Index Numbers and Aggregation; Leading indicators ; Neural Networks and Related Topics ; Forecasting Models; Simulation Methods
• Subject: trading strategy ; stock market index ; neural networks ; cointegration ; Börsenkurs ; Aktienindex ; Neuronale Netze ; Prognoseverfahren ; Theorie

Contributors, Places and Time

• Event: Geistige Schöpfung
• (who): Haefke, Christian ; Helmenstein, Christian
• Event: Veröffentlichung
• (who): Institute for Advanced Studies (IHS)
• (where): Vienna
• (when): 1995

Further information

• Handle: http://hdl.handle.net/10419/68636
• Last update: 10.03.2025, 11:43 AM CET