This paper suggests the use of a forecasting model that utilizes the Markov chain when conditions show that there is insufficient time-series data. The Markov chain corresponds to the vector auto-regressive (VAR) model of the first order. If the transition probability matrix can be appropriately estimated, the forecasting model using the Markov chain can still be constructed with less time-series data even though the amount of data is sufficient to estimate the VAR model. The forecasting model, used at two points in this paper, estimates the transition probability matrix based on industrial macro data in Japan. The use of this application at a regional level will also be discussed along with changes that could possibly affect the future of industrial structure in Japan.