Economic and Statistical Modelling of Production Processes in the Regions of the Arctic Zone of the Russian Federation

The study of economic processes in the regions of the Arctic Zone of the Russian Federation (AZRF) using economic-statistical modelling methods is an important area of Arctic research due to the possibility of displaying statistically significant relationships of economic processes and phenomena and forecasting economic dynamics, but the potential of such modelling is limited by the specificity of the processes of a number of regional economies in the Arctic, distorted by the increased state presence and active man-agement intervention in the Arctic. Therefore, economic and statistical studies of the Russian Arctic regions are rare in relation to the popularity of the Arctic topics in regional studies. The aim of the study is economic-statistical modelling of production processes in the regions of the AZRF using the production function (PF) toolkit. At the first stage, the analysis of correlations between GRP and factors of production was carried out, on the basis of which the AZRF regions were divided into three groups: 1) regions in which the mutual behavior of the main factors of production (labor and capital) fits into the generally accepted concepts (the Russian Federation as a whole, the Nenets Autonomous Okrug, the Yamalo-Nenets Autonomous Okrug, the Republic of Sakha (Yakutia)); 2) regions in which the mutual behavior of the main factors of production does not fit into the classical concepts: there is a sufficiently strong positive relationship with only one of the factors of production (the Arkhangelsk Oblast, the Krasnoyarsk Krai, the Komi Republic); 3) regions for which there is no sufficiently strong positive relationship between GRP and factors of production (the Murmansk Oblast, the Republic of Karelia, the Chukotka Autonomous Okrug). At the second stage, at least 4 PF models were built for the regions of group 1 (the best model was selected using the Akaike information criterion adjusted for small samples). For the regions of group 2, Cobb-Douglas PF models and single-factor models were built, in which a factor with a positive relationship with GRP was included. The construction of PF models for the regions of group 3 is impossible.