Unit Roots and the Demand for Cigarettes in Turkey: Pitfalls and Possibilities

Abstract

The partial adjustment model for cigarette demand in Tansel (1993) is formulated as a restriction on the more general VAR model. The question whether the Tansel estimation results are spurious as claimed by Cameron and Collins (1998) is addressed in this framework. The role of intervention dummies and data transformations for the cointegration results is discussed in some detail. Finally, a more general model is estimated and two steady-state relations measuring demand and supply behaviour in the cigarette market are identified.