Time Varying VAR Analysis for Disaggregated Exchange Rate Pass-through in Tunisia

Our paper follows the "Time Varying Parameter VAR with Stochastic Volatility" (TVP VAR) approach developed by Primiceri (2005): Bayesian estimation with time varying coefficients and stochastic volatility. Our paper contributes to the literature by examining if the impact of monetary and exchange rate shocks have varied over time in Tunisia through a disaggregated analysis of exchange rate pass-through by introducing time variability in two ways; firstly, by assuming That all the coefficients of the VAR model are variant in time, and secondly, in the temporal variance-covariance matrix, that is the error term’s volatility of the VAR model. The multivariate stochastic volatility aims at capturing the heteroskedasticity of shocks and non linearities in the simultaneous relationships between the variables of the model. In fact, it allows us to capture abrupt and progressive changes in state variables. Given the structural and institutional changes in the Tunisian economy over the last few decades, it is important to emphasize the possibility of such a temporal variation in the empirical methodology. To the best of our knowledge, this work is among the first to apply the TVP-VAR approach with stochastic volatility to the shocks of monetary and exchange rate policies in Tunisia. Overall, the findings confirm that the modeling approch; i.e the TVP-VAR, is the best tool to analyze the impact of these shocks in Tunisia. The results of the study can help the short- and long-term decision-makers in Tunisia to adopt appropriate strategies for conducting monetary policy as well as containing inflation.