Abstract:

The focus of the present paper is forecasting in economics, finance, and management. Forecasting is the process of making rational statements about future events or the state of unexamined entities, based on observation and the use of knowledge about the regularities that characterize the forecasted phenomenon and the knowledge of relationships between it and other phenomena. Forecasting is the initial component of the triad: forecasting, goals, and planning. This triad is an integral part of any management decision-making process and each part of it serves to reduce the risk in an enterprise. All organizations, from international communities to companies, undertake activities that require forecasting in their operations.

During the recent COVID-19 pandemic, many pre-existing and studied regularities have ceased to function. The paper discusses the problem of applying forecasting methods used in times of stability to forecasting in times of instability. Lack of stability means an increase in the uncertainty of real processes, and thus an increase in the uncertainty of the forecasting model. This increase can be attributed partial uncertainties: random disturbances, parameters, and data.

The macroeconomic forecasting model is the subject of analysis. Economic forecasting is hierarchical, therefore, in order to manage processes on a micro-scale, it is necessary to predict the state of processes on a macro-scale. The aim of this paper is to propose a forecasting model with greater, both in construction and method, robustness to uncertainty, especially during its increase due to the pandemic. I suggest using business survey data and a statistical regression variable selection method, in a recursively structured model. The model also allows for generating forecasts without having to make assumptions regarding regressor values during the forecast period.