Abstract

This paper systematizes the experience of empirical research on
news shocks, in particular on models of vector autoregressions, as well
as dynamic stochastic models of general equilibrium. The review examines the work
with various types of shocks, presents the main disadvantages of different classes of models
used to analyze the impact of news shocks on developed and developing
economies. The main prerequisites that need to be included in the DSGE are highlighted
models so that these models can demonstrate a co-directional change
the main macroeconomic indicators in response to the news shock. It has also
been demonstrated that VAR models and DSGE models use many similar ideas –
it is useful to include predictive indicators in each of these models. During the research
it was found that the proportion of explained variation in works with DSGE models is very
different from the proportion of explained variation in works with VAR models. A possible
reason may be that VAR models are characterized by the problem
of non-fundamentality.