Nonlinear dynamics of U.S. equity factor portfolios

Area of Science:

• Quantitative Finance
• Financial Econometrics
• Chaos Theory

Background:

• U.S. equity factors, such as value, size, momentum, and reversal strategies, are widely studied.
• Traditional financial models often assume efficient markets and random security returns.
• Behavioral anomalies and risk premia suggest deviations from purely random price movements.

Purpose of the Study:

• To investigate the presence of nonlinear dependence and chaotic dynamics in U.S. equity factor returns.
• To test the validity of the random walk model against evidence of complex system behavior.
• To provide novel insights into the underlying mechanisms driving financial markets.

Main Methods:

• Analysis of daily return data for key U.S. equity factors (value, size, momentum, long-term reversal, short-term reversal) from the early 20th century.
• Application of the Brock-Dechert-Scheinkman (BDS) test to detect nonlinear dependence.
• Calculation of chaos indicators: maximal Lyapunov exponent, correlation dimension, and Kolmogorov entropy.

Main Results:

• The Brock-Dechert-Scheinkman test provided statistically significant evidence of latent nonlinear dependence in equity factor returns.
• Measures of maximal Lyapunov exponent, correlation dimension, and Kolmogorov entropy indicated patterns consistent with chaotic behavior.
• These findings contradict the predictions of the standard random walk model for security returns.

Conclusions:

• U.S. equity factor returns exhibit complex dynamics characterized by nonlinear dependence and chaos.
• The random walk model is insufficient to explain the observed behavior of these financial assets.
• The presence of chaos suggests potential predictability and offers new avenues for financial modeling and risk management.