Adjusting For The Chinese New Year: An Operational Approach

In macro surveillance work, we often find it difficult to assess the state of the Mainland economy based on unadjusted monthly or quarterly data as the underlying economic trends are obscured by seasonal variations and the effect of moving holidays. Standard procedures exist for seasonal adjustment, but they do not deal with unusual data movements caused by the Chinese New Year (CNY) ¡V a moving holiday in the Gregorian calendar. In this study we refine seasonal adjustment by taking into account the CNY effect. Two methods are used to pre-adjust a series before the actual seasonal adjustment step: a) taking the average of January and February, and b) using CNY dummies. The empirical analysis shows that taking into account the CNY effects noticeably improves the quality of seasonal adjustment. There is, however, no clear winner between the two techniques for adjusting for the CNY effect. Given the relative operational ease of averaging January and February data, we intend to adopt this method in our regular data analysis.