An Application and Extension of the Markov Chain Model to Cereal Production

Given aggregate time‐series data, conventional Markov Chain estimation of the stationary transition probabilities is performed for a model of the cereal market in Great Britain. The possibility of non‐stationary probabilities is examined in the light of these initial results, and the basic model developed to allow exogenous variables to enter via their influence on these parameters. The evidence from the paper would appear to suggest that, at least in the case considered, time‐varying transition probabilities are a most attractive alternative.