Improving the Prediction Accuracy of the Integral Indicators by the Means of Combining Forecasts

• Published in: Finance: Theory and Practice Vol. 21; no. 5; pp. 118 - 127
• Main Authors: Frenkel, Alexander A., Volkova, Natalia N., Surkov, Anton A.
• Format: Journal Article
• Language: English; Russian
• Published: Government of the Russian Federation, Financial University 01.11.2017
• Subjects: интегральный индекс; объединение прогнозов; поворотные точки; темп роста; экономическая динамика; экономическое прогнозирование
• ISSN: 2587-5671; 2587-7089
• DOI: 10.26794/2587-5671-2017-21-5-118-127

Topic. If we need to predict the future economic development of the state it is necessary to build indicators that could be detectors of economic development. These detectors are integral indices that can describe the overall state of the economy of the state and can warn of turning points in the development in the future. The paper discusses methods of constructing such integral indices and compares them with the rates of industrial production. We provide analysis how to improve the prediction accuracy of the integrated indices through the use of methods of combining forecasts. Combining forecasts proved to be in practice an adequate method of improving the accuracy of forecasting in conditions of uncertainty of choice between individual forecasts.Purpose. The purpose of this work was the construction of three integrated indices describing the general state of the Russian economy: leading, coincident, and lagging, their statistical analysis, calculation of forecast values of the considered indices and the estimation of the influence on prediction accuracy of combining forecasts.Methodology. The study used statistical methods to construct the integrated indices as well as statistical methods of forecasting and the technique of building of combining forecasts.Results. The results of our researches have become integral indices for the Russian economy in the period from 1999 to 2016, and their statistical comparison with observed rates of industrial production. This created an opportunity for making the forecast of development of Russian economy for the next year and comparing the forecast results with the actual data for the first four months of 2017. There are built several prediction models which were combining into the overall forecast. Combining forecasts have improved the prediction accuracy.Conclusions. The result of the work allows concluding that the combining forecasts substantially improves forecasting accuracy of integrated indices and allows using the technique of amalgamated forecasts to predict “turning points” in economic development.