우리나라 휘발유시장 간의 가격발견과 동태적 상호의존성에 관한 연구

We employ a vector error correction model (VECM) and directed acyclic graphs (DAG) to investigate the price discovery processes of the gasoline prices in 7 major cities in Korea. We also analyze dynamic interdependences of the gasoline prices. Data running from July 3, 2008 to July 3, 2011 consist of daily prices of medium level gasoline. Empirical results from VECM and DAG show that the 7 gasoline prices are cointegrated and thus the gasoline markets tend to be integrated in the long run. The variance decomposition indicates the price discovery processes rely on the length of time-horizon. Moreover, DAG anlaysis shows that Daejeon which is the most competitive gasoline market plays an important role in the flow of contemporaneous price information while, according to the variance decomposition, Ulsan in which the biggest petroleum refinery in Korea is located is an important market in the price discovery process as the time-horizon increases.