Riassunto analitico
Using a new frequency domain structural VAR identification technique, Angeletos et al. (2020) find that a single shock, which they call the main business cycle shock (MBC), can account for the bulk of macroeconomic fluctuations. This finding challenges the widespread view in the economic theory that two shocks are the main driving forces of the economic cycle, where these shocks are an inflationary demand shock and deflationflary shock. Following this path, Forni et al. (2021), using the same frequency domain SVAR methodology, find results which are consistent with the mainstream theory of two shocks. However, SVAR analysis suffers from a possible information problem known as non-fundamentalness: the information embedded in the VAR might not be enough to recover the structural shocks. This is the reason why VAR results are often very sensitive to the choice of the variables. This problem can be solved using a factor model approach appropriately adapted to SVAR analysis: by doing so, one can virtually include all the macroeconomic information. In this thesis, we first estimate a structural dynamic factor model, then we identify the shocks in the frequency domain that have the maximum contribution to the explained variance of a targeted variable over a targeted frequency band, in the spirit of Angeletos et al. (2020) and Forni et al. (2021). Using a large data set containing 215 variables, referred to quarterly U.S. data from 1960:Q1 to 2020:Q2, we find a clear-cut result: the
business cycle macroeconomic activity is best captured by two shocks, a permanent shock that resembles a supply shock and a transitory shock that resembles a demand shock. We also find that the long run macroeconomic activity can be explained by just one shock.
|
