Predictable variation and profitable trading of US equities: a trading simulation using neural networks

Abstract A switching rule conditioned on out-of-sample one-step-ahead predictions of returns is used to establish investment positions in either stocks or Treasury bills. The economic significance of any discernible patterns of predictability is assessed by incorporating transaction costs in the simulated trading strategies. We find that ANN models produce switching signals that could have been exploited by investors in an out-of-sample context to achieve superior cumulative and risk-adjusted returns when compared to either regression or a simple buy-and-hold strategy in the market indices. The robustness of these results across a large number of stock market indices is encouraging. Scope and purpose A large body of evidence has accumulated suggesting that stock returns are predictable by means of publicly available information on a number of financial and macroeconomic variables with an important business cycle component. Previous research has, for the most part, relied on standard statistical techniques (e.g., regression analysis) with unduly restrictive assumptions presumed to hold in the underlying data-generating process. This paper reexamines the evidence regarding predictable variation in US stock returns using both artificial neural network (ANN) and regression, and compares simulated trading results obtained from ANN models with those obtained from regression.

[1]  B. LeBaron,et al.  Nonlinear Dynamics and Stock Returns , 2021, Cycles and Chaos in Economic Equilibrium.

[2]  W. Ferson.  Changes in Expected Security Returns, Risk, and the Level of Interest Rates , 1989 .

[3]  L. Glosten,et al.  Economic Significance of Predictable Variations in Stock Index Returns , 1989 .

[4]  Douglas M. Patterson,et al.  Evidence of Nonlinearity in Daily Stock Returns , 1985 .

[5]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[6]  Allan Timmermann,et al.  Dangers of Data-Driven Inference: The Case of Calendar Effects in Stock Returns , 1998 .

[7]  H. White.  Some Asymptotic Results for Learning in Single Hidden-Layer Feedforward Network Models , 1989 .

[8]  Efraim Turban,et al.  Expert systems and applied artificial intelligence , 1992 .

[9]  E. Fama,et al.  Dividend yields and expected stock returns , 1988 .

[10]  G. William Schwert,et al.  Asset returns and inflation , 1977 .

[11]  Christian Lebiere,et al.  The Cascade-Correlation Learning Architecture , 1989, NIPS.

[12]  E. Fama,et al.  BUSINESS CONDITIONS AND EXPECTED RETURNS ON STOCKS AND BONDS , 1989 .

[13]  Kazuo Asakawa,et al.  Stock market prediction system with modular neural networks , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[14]  M. Hashem Pesaran,et al.  A Simple Nonparametric Test of Predictive Performance , 1992 .

[15]  A. Timmermann,et al.  Predictability of Stock Returns: Robustness and Economic Significance , 1995 .

[16]  H. White,et al.  Data‐Snooping, Technical Trading Rule Performance, and the Bootstrap , 1999 .

[17]  M. C. Jensen.  Some Anomalous Evidence Regarding Market Efficiency , 1978 .

[18]  G. Schwert,et al.  Stock Returns and Real Activity: a Century of Evidence , 1990 .

[19]  S. Dutta,et al.  Bond rating: a nonconservative application of neural networks , 1988, IEEE 1988 International Conference on Neural Networks.

[20]  Edward F. Renshaw,et al.  Modeling the Stock Market for Forecasting Purposes , 1993 .

[21]  Donald B. Keim,et al.  Predicting returns in the stock and bond markets , 1986 .

[22]  J. Campbell.  Stock Returns and the Term Structure , 1985 .