Abstract
The integration properties of Turkish velocity series are investigated by employing recently developed procedures (Zivot and Andrews (1992) and Perron (Journal of Econometrics, 80, 355–85, 1997)) which allows stationarity around an endogenously estimated structural break point under the alternative hypothesis. The evidence suggests that real currency balances and currency velocity are stationary around a broken trend. The estimated break point coincides with a major policy regime change in Turkey. Broad money velocity and real broad money balances are found to be nonstationary even after allowance for a broken mean and trend.