Construction of an integrated regression-extrapolation model for forecasting operating costs of railroad station infrastructure
Abstract
This study investigates the process that drives operating costs for maintaining the station infrastructure of the Ukrainian railroad. Financial support to the country’s railroad facilities requires reliable cost forecasts. However, the inflation of 2014 and the beginning of full-scale military aggression against Ukraine in 2022 significantly changed the economic conditions, thereby compromising the accuracy of conventional methods for substantiating investment decisions and budgets. To solve this problem, an integrated regression-extrapolation model has been built. Based on the results of the correlation-regression analysis, two statistically significant factors were left out of several considered factors. To take into account the impact of the crisis years, it was proposed to introduce dummy variables, which made it possible to separate their impact and obtain an average approximation error of 3.66%. The constructed model explains more than 98% of the variation in costs; the residuals do not have systematic deviations; and all factors are statistically significant. According to the Consensus Forecast by the Ministry of Economy, three scenarios for predicting specific operating costs for maintaining station infrastructure over 2025–2027 were devised. The optimistic scenario is built on the minimum values of the selected factors and gives the lower limit of costs under favorable development of events. The pessimistic scenario, formed on the maximum values, is a guideline for assessing financial risks. The basic forecast corresponds to the average values and is the basis for budget formation. The resulting forecasts could be used for budget planning, risk assessment, and justification of investments in railroad infrastructure. When updated macroeconomic forecasts appear, the model makes it possible to quickly calculate the scenario values of costs, which makes it convenient for annual use in the planning cycle