A Non-Causal Autoregression Based Estimation of the New Keynesian Phillips Curve for the Turkish Economy

Abstract

This study investigates the dynamics of inflation for Turkish economy from 1990 to until 2013 under the New Keynesian Philips Curve framework by using the non-causal time series methodology developed by Lanne and Saikkonen (2011) and Lanne and Luoto (2013). There is a vigorous ongoing debate on the appropriate specification of the New Keynesian Phillips Curve. A non-causal auto regression based estimation of the New Keynesian Phillips Curve approach enables us to avoid both potential weaknesses related to the Generalized Method of Moments estimation and a number of problems arise from the determination of marginal cost proxy without considering controversial ways in the literature. Our findings indicate that Hybrid New Keynesian Phillips Curve is valid for Turkish economy for both 1990s and 2000s and the role of forward looking behaviour has increased after implementing Inflation Targeting. Based on our findings, we inferred that the CBRT has become more effective for managing inflation expectations and monetary policy after adopting inflation targeting. However, inflation dynamics of the Turkish economy has still included inherited forms even under this regime. We indicated that this result has arisen from the failure of the CBRT to keep its inflation targets.