|
Abstract:
|
This study analyzes several variables using a statistical model to represent the relationship between these quantities as they change over time. A multivariate technique, the Vector Autoregressive Model (VAR), predicts the alliance between these variables. Moreover, the Structural Vector Autoregressive (SVAR) model is used to impose restrictions on the residual covariance matrices that explain structural shocks from the reduced form of VAR. For a better explanation, the Bayesian Autoregressive model (BVAR) is used. The study has the following objectives. First, normalizing the data and checking for stationarity to determine the persistence of the model. Second, building three models through the lag selection, testing structural breaks in the residuals, and conducting Granger causality tests for linear or non-linear granger cause. Third, finding the impulse response function, evaluating the shock for each variable on others, and estimating variance decomposition to see the lag that is accountable for variability. The primary purpose of this study is to find out the most suitable model.
|
