Low-frequency relationship between money growth and inflation in Turkey

Abstract

This paper examines the long-run and medium-run predictive relationship between money growth and inflation in Turkey for the period 1986m1-2018m12, using frequency-domain methods. For the full sample, the measures of spectral coherence and gain spectrum suggest a one-to-one relationship, and the frequency domain decomposition of the Granger causality test indicates a bidirectional predictive relationship between the two variables at zero frequency. As suggested by the wavelet coherence, we also analyzed the two subperiods before and after 2006 separately. Our results suggest that while both variables have predictive power for each other in the second subperiod, only money growth helps predict inflation in the first one. In order to prevent spurious results, the analysis is rerun in a multivariate Vector Autoregression (VAR) system, where output growth, interest rate, exchange rate growth, and domestic debt growth are included as additional variables. We observe that while money growth has predictive power for inflation in the first subperiod, this relationship disappears in the second one. We argue that the change in the relationship between the two variables at low frequencies after 2006 is primarily a result of the decrease in fiscal dominance of the government, the CBRT's switch to the inflation targeting regime, and the CBRT's unconventional monetary policy framework.