An empirical test of the Causative Markov model of consumer behavior

This paper examines an extension of the Markov model of consumer brand switching behavior, which captures the nonstationarity in the consumer decision processes over time. The extended model has been labeled as the Causative Matrix model. The model is put to test on empirical data, and compared with the stationary version of the model. The Causative Matrix model is shown to be a better forecasting tool. Its ability to yield additional insights into consumer brand choice processes is also discussed.