A thresold regression estimation of philips curve: Turkey case

Abstract

In this paper, using the threshold regression model developed by Hansen (1996, 2000), we estimated an extended Phillips curve relationship with variables playing an important role in the formation of inflation. We also used Hamilton's (2001) flexible nonlinear inference approach to test whether nonlinearity existed and if the threshold model was appropriate. According to our findings, the model in which the estimated threshold value for capacity utilization rate is equal to 75% is quite reasonable for modeling the Turkish Phillips curve. This result supports the idea that high and low capacity measures may have different impacts on inflation. Moreover, unlike other studies in Turkish Phillips Curve literature, it presents a stable nonlinear relationship between the real economic activity and inflation in Turkish economy.