Indicators for assessing the risk of execution of state contracts with a long life cycle

Abstract

The subject of the study is the process of monitoring the execution and assessing the risk of execution of government contracts in the Russian Federation. The research methodology consists of studying statistical data on government procurement in Russia; enriching data from the UIS with additional data from market analysis systems; analyzing the subject area and identifying potentially valuable categorical data that have not been previously studied by other scientists and applying the tools of one-way analysis of variance to these data in order to assess their statistical significance for solving problems of predicting the execution of government contracts. By applying the ANOVA method to a dataset that included more than 83 thousand consolidated records, results were obtained confirming the significance of a number of categorical features relating to the supplier’s industry, its legal form and region for predicting the execution of government contracts. At the same time, it was revealed that such characteristics as form of ownership, method of placing an order, legal category of the size of the supplier’s business and budget level are not statistically significant for forecasting purposes. The results obtained can be used by researchers in the course of cluster analysis, exploratory data analysis, and when constructing an ensemble of models for predicting the execution of government contracts. The results obtained expand and deepen existing approaches in terms of searching for new significant features based on information capabilities contained in government information systems and other big data sources.