In search of Q: results on identification in structural vector autoregressive models

The thesis consists of three independent research papers in the area of identification of structural vector autoregressive (SVAR) models. The first paper studies the government spending shocks on output. New sign restrictions on the impulse responses are introduced to the fiscal policy literature. Upon imposing the new sign restrictions, we find contractionary output reactions to positive government spending shocks. The results are robust to several alternative model and sign restriction specifications. The second paper investigates the implications of including more (forward-looking) information in a classical monetary policy SVAR. We augmenting the classical specification with federal funds futures and factors extracted from a large macroeconomic dataset. It is established that information-augmentation does not necessarily yield shocks that are correlated more strongly with benchmark monetary policy shocks. The empirical conclusions based on the baseline and augmented specifications are similar: monetary policy shocks contribute only marginally to the evolution of (real) macroeconomic variables. The third paper‘s empirical motivation is to quantify the effects of monetary policy shocks on asset prices. We set up a monetary SVAR including the S&P500 series, and we use set identifying sign and zero restrictions. It is established that the majority of monetary policy shocks estimated from admissible structural models correlate only weakly with benchmark monetary policy shocks. Structural models with medium correlations, imply impulse responses that vary widely in their shapes and magnitudes. Concentrating, however, only on the 100 models with the highest correlations uncovers negative, but near-zero asset price responses to positive monetary policy shocks, coupled with mildly positive output responses.