PREDICTED YIELDS MODELING AND THEIR USING IN PROBLEMS OF EFFECTIVENESS ESTIMATION FOR INVESTMENT PORTFOLIOS

The approach to the modeling the assets prices has been suggested in the work and it permits to receive high imitation and forecasting properties of yields. Using forecasting yields while forming the real investment portfolio enables to form the portfolio more efficiently and to receive more profit as compared to the traditional approach.In case of econometric modeling of the asset price we proceeded from the fact that there are observed the effects of continuous and discrete changes in the price development. That's why the model itself and the procedure of its construction had to provide for the parameters of the continuous model and the discrete effects were identified iterative by, and in this way the interaction of the continuous and the discrete components configured the regress and for the adequate reflection of the real process. Then continuous-discrete model with two-digit dummy variables has been used for the construction of variants of predicted yields, and the logit-model has been used to determine the optimal variant.The approbation of the developed algorithm has been performed using real statistical data of stock prices for six Ukrainian companies. High imitation and forecasting properties of the obtained models point to possibility of using the results during modeling real economic systems.