Abstract
The objective of this study is to apply the control theory to the identification and forecasting problems in economic systems, and to test its validity. In this paper, a macroscopic economic system of Japan is considered.
The modeling procedure consists of three stages. In the first stage, the structure is determined so that economic policy can be taken into account in the model. The system behavior is described as a perturbed behavior along the nominal trajectory so as to obtain a linearized model. In the second stage, parameters are estimated from the actual observation data. And, in the last stage, the behavior of this system is forecasted by using the Kalman filter and predictor.
The results of the forecast show good agreement with the actual behavior of this system.
