DOUBLE-EDGED SWORD: TURKISH UNEMPLOYMENT UNDER INFLATION TARGETING A Structural and Bayesian Vector Autoregression Analysis

Abstract

This paper looks at the effect of inflation targeting (IT) on Turkey’s inflation andunemployment rates, focusing on whether the inflation and unemployment rates becamemore susceptible to internal and external shocks after the adoption of IT by the CentralBank of the Republic of Turkey (CBRT). For this purpose, we use a four-variable structuralvector autoregression (VAR) model, adopted from King and Watson (1994) and Djivre andRibon (2003), and then adapt the structural model to a Bayesian VAR setting to utilizeBayesian econometric inference methods to answer the research question. Through thegeneration of cumulative impulse – response functions, which show the impact ofexogenous shocks to the macroeconomic variables, and forecast error variancedecompositions, which depict the pass-through of the shocks to one variable to the others,we conclude that the unemployment rate became more susceptible to supply and demandshocks after the official adoption of inflation targeting in 2006. Results from the BayesianVAR model also indicate that the CBRT became less reactionary to supply and demandshocks after the adoption of inflation targeting, a finding that is corroborated by the resultsfrom the structural VAR model. Overall, we find that the adoption of inflation targetingmay have come at the cost of rising unemployment, although caveats need to be attachedto this policy conclusion, not least because of a short sample period and the use of aMinnesota prior with a large tightness hyperparameter.