APPLICATION OF VECTOR AUTOREGRESSIONS FOR FORECASTING MONETARY POLICY

Abstract

The purpose of the study is to consider the theoretical and empirical application of methods and models of vector autoregressions to analyze the infl uence of various macroeconomic variables in the construction of a monetary policy model.Methodology. The research methods used are generalization of experience regarding the use of vector autoregressions, factor analysis, methodology for evaluating VAR models containing fi fteen real, price, monetary and external variables. A number of tests were conducted to assess the quality of the analyzed model: impulse response analysis, forecasting and simulations.This article analyzes the infl uence of factors on each other, as well as the interpretation of the results, which can be further used to obtain practical recommendations for improving the methods of research and forecasting monetary policy.Originality / value of the research. The paper analyzes the advantages and disadvantages of diff erent approaches in the construction of vector autoregressive models, both in the selection of factors and their preparation for use in the model. This article examines the observation period from 2010 to 2021, that is, before and after the introduction of the infl ation targeting regime, and the assessment of the pandemic shock in Kazakhstan, without aff ecting the shocks of 2022.Findings. The work carried out made it possible to verify the applicability of vector autoregression methods, this statement is confi rmed by the inverse predictive power of the models. In this paper, the eff ectiveness of the proposed models was evaluated on macro factors in Kazakhstan.