Forecasting in a Dynamic Stochastic General Equilibrium Model for Turkey: A Bayesian Approach

Abstract

In this study, a small open economy theoretical model based on Galí and Monacelli (2005) which is a version  of dynamic stochastic general equilibrium (DSGE) model will be combined with a vector autoregression (VAR) model and the resulting DSGE-VAR model will be used to make long-term projections through Bayesian estimations for the Turkish economy. DSGE models are generally used in the analyses of short-term macroeconomic policy within the New Keynesian framework. The long-term projections are made generally via the VAR models. In this study, the two approaches will be combined, and projections will be made using a VAR model starting from a structural DSGE model for Turkey. Because in literature, there are not many long-term projections and estimations made within such a framework, this study will contribute the literature. The estimation and forecasting results are obtained for Turkey’s 2023 gross domestic product (GDP) and per capita GDP. The results show that those 2023 targets will not be met. So if Turkey is serious and insistent on these targets, this study may be warning to policymakers and the public to take the necessary measures when there is still enough time and opportunity.

Keywords: Dynamic Stochastic General Equilibrium-Vector Autoregression; Small Open Economy; Bayesian Methods; Projection

Australian Academy of Business and Economics Review, vol 3, issue 1, January 2017, pp 1-12