VAR model

USING GRETL (or other software if possible):
1. Suppose to evaluate, through a structural VAR model, the dynamic impact of oil price shocks on some
macroeconomic variables in a “small open economy”. Let consider the following vector of variables yt =
(oilt , inft ,growtht)’, where oilt is the oil price (in domestic currency), inft is the inflation rate (growth rate of
the consumer price index), and growtht is the growth rate of the gross domestic product. Using the first
three series reported in the dataset, specify and estimate a VAR model and discuss the propagation
mechanism of shocks by calculating structural impulse response functions.
Let choose one of the three variables and estimate an appropriate ARMA model, commenting on the
differences with respect to the corresponding equation in the VAR model.

2. The Purchasing Power Parity (PPP) states that once converted to a common currency, national price
levels should equalize. From a theoretical point of view, if PA and PB are two price indices related to
Country A and Country B and EAB is the nominal exchange rate, the PPP relation can be written as
Or, when using a log transformation,
where lower-case letters denote logs of the variables. From an empirical point of view, if the series are
nonstationary, one possibility to test for this condition is to refer to the notion of cointegration. Using the
series reported in the dataset (already transformed in logarithmic terms), verify first that the variables are
nonstationary, and then provide an empirical analysis to confirm the PPP hypothesis. Moreover, through an
error correction model, verify the reaction of the three variables to disequilibria in a “possible”
cointegrating relation.






Sample Solution