In this paper we developed an empirical small open economy model for macroeconomic projection of the Israeli economy. The model is estimated for the period between 1997Q3 and 2008Q3. Our primary purpose is inflation forecasting and the evaluation of the effect of monetary policy on prices. The endogenous variables in the model are: growth of business sector product, the nominal effective depreciation of the Shekel, the CPI inflation rate, and the Bank of Israel policy interest rate. Our empirical strategy constitutes a compromise between "letting the data speak" and imposing structural constraints on the coefficients of the model - as suggested by standard New-Keynesian theory. To that end we estimated four models: a simple unrestricted VAR (Vector Auto-Regression), a model with selective exclusion restrictions on the exogenous variables in each equation, a model with structural constraints (in addition to the exclusion restrictions), and finally the structural model is augmented by rational expectations. The performance of the model improved, both in term of forecasting ability and in terms of its dynamic properties (as reflected by the Impulse Response Functions), with the imposition of structural constraints. However, following the imposition of rational expectations its forecasting performance deteriorated, leaving the model with structural constraints as the version generating the best results. Finally, analyzing the dynamic responses of the model to structural shocks, we conclude that the imposition of additional long-run restrictions is required.

For a PDF of the full article