At first, this study examines the relevance of learning effects, land improvement, degree of agricultural engagement and production adjustment (reduction of cultivated area) policy with the production of Japanese rice. Secondly, the learning elasticity and learning rate of Japanese rice are estimated. As the results, learning effects to the production of Japanese rice is presented. Then, the equation of convergence of “the initially lower prefectures of rice crop yield per 10a have indeed grown faster” is analyzed, and the cross-sectional coefficient of variation of rice crop yield per 10a also declined between 1970-90. Finally, the empirical evidence shows that the speed of convergence for rice crop yield per 10a is affected by the learning effects.
In this paper, a new extraction method of the variable seasonal index was developed which based on learning the extraction methods of the fixed and the variable seasonal index. The new method calls Link Relative Moving (LRM) method. We tested that the variable seasonal index of LRM method used Monte Carlo Experiment, and the result was compared the Economic Planning Agency (EPA) method. We studied the influence of the different seasonal adjustment index on cyclical fluctuation and the irregular fluctuation. We evaluated that total effect of the seasonal index on LRM method, EPA method and Link Relative method. After decision of period by power-spectral analysis, we compared the irregular fluctuation results on LRM method, EPA method and Link Relative method. In conclusions are; firstly, LRM method is a valuable method among extraction methods of the variable seasonal index; secondly, the result of Monte Carlo Experiment is stabilized in seasonal index of LMR and EPA methods, and the extract power of seasonal index is stronger in LMR method than EPA method; thirdly, the irregular fluctuation is smaller in seasonal adjustment index of LMR method than EPA method and Link Relative method.