City Price Convergence in Turkey with Structural Breaks

This paper explores relative price convergence for 18 cities in Turkey. The convergence implies stationarity in the long run. Henceforth, to observe whether price convergence occurs or not, this study conducts unit root tests following Lee and Strazicich (2003) with two structural breaks in level and/or trend. The test statistics reveal that 13 out of 18 consumer price indexes converge. The half-life measurement points out that the speed of convergence of each city is considerably high. This result indicates that the half of the cumulative shocks persists for a short time period. The contribution of this work lies in three points. First, this is the first work which employs relative individual city prices in Turkey through convergence analyses. Since, unfortunately, the Turkish Statistical Institution does not continue to yield current individual city price indexes, this work needed to launch the previous decades' data points to estimate convergence parameters. Secondly, this paper aims at providing policy makers with some ex-post predictions about individual city prices. This, in turn, will help policy makers in conducting ex-pots and ex-ante forecast analyses about city level, and/or, regional level, and/or, national level prices of Turkey. Third, this work aims at, as well, providing researchers with efficient and consistent convergence estimation models. To this end, this paper suggests that researchers follow the methodologies of Lagrange multiplier (LM) unit root and half-life analyses simultaneously in determining the optimization paths through iterations, since LM estimator(s) might be thought still considerably efficient, consistent and unbiased estimator(s) among other available ones in the literature in terms of 2016. This later issue needs to be discussed, of course, in detail, in future works through further linear and nonlinear power tests.