Forecasting Models for the Economic Security of Regional Clusters of the National Economy

Abstract

Introduction. The principal internal factors of the economic security of regions are defined by principal component analysis and their correlation with respective indicators is determined. Four principal factors (components) of the economic security of regions are identified. Forecasting models for the economic security of regions of the national economy are constructed by regional cluster using the results of the study. The study is performed by data on main socio-economic indicators of 24 Ukrainian regions over 2008–2015 and by economic security clusters (classified as “good”, “sufficient” and “satisfactory”) defined in previous studies.
 The purpose of the study is to construct sets of forecasting models for the economic security of the national economy by regional cluster.
 Methods. Methods of theoretical generalization are used to study and systematize the factors of the economic security of regions; multidimensional statistical methods such as principal component analysis are used to define their correlation with the respective indicators; the method of comparison allows for defining common and distinctive features of the results of computation of the economic security of regions using the constructed models.
 Results. Principal component analysis used for defining the internal factors of economic security of regions and their correlations with the respective indicators allows for identifying four principal factors (components) of the economic security of regions. Results of principle component analysis show that the significance of the first four latent factors for the overall sample is 86.86% of the dispersion of the analyzed variables (13.14% error); for “good” cluster it is 87.3% (12.7% error); for “sufficient” cluster it is 88.03% (11.97% error); for “satisfactory” cluster it is 86.0% (14.0% error). Four sets of forecasting models are constructed, each consisting of a mathematical description of four principal components for each sample. The implementation of the constructed models confirms the inclusion of the regions to the respective regional clusters. The dynamics of the indicators shows the economic situation and processes in the regions in the respective periods. 
 Conclusion. The applied character of the constructed models is substantiated by their practical implementation using actual socio-economic data for the regions represented in each of the regional clusters (Dnipropetrovsk, Vinnytsia, and Volyn regions), over the period of 2008–2015. Further studies should focus on making the scale of the economic security performance for the constructed forecasting models, to determine the performance and estimate the probability of insecurity occurrences in the economy using expected (planned) figures of economic development in a region.