An investigation of structural breaks on spot and futures crude palm oil returns

• Main Authors: Zainudin, Rozaimah, Shaharudin, Roselee Shah
• Format: Article
• Published: American-Eurasian Network for Scientific Information, Jordan 2011
• Subjects: HF Commerce
• Online Access: http://eprints.um.edu.my/23011/; http://www.ajbasweb.com/old/ajbas/2011/September-2011/1872-1885.pdf
• Institution: Universiti Malaya

The implications of omitting structural break in volatility clustering modelling process are largely discussed in various developed macroeconomic and finance variables. The empirical evidence highlights the importance of this structural break in modelling process that gives a more accurate volatility persistency estimation result. The main purpose of the paper is (i) to identify the existence of structural breaks in crude palm oil return and (ii) to demonstrate the implication of structural break in volatility clustering estimation process. This paper tests for the presence of structural breaks in spot and future crude palm oil returns. Using daily data that span from January 1996 to August 2008, we test the existence of structural changes in tested crude palm oil series mean via the Bai and Perron procedure. Using the Inclan and Tioa Iterated Cumulative Sums of Squares (ICSS) algorithm procedures, we proceed to identify any structural changes in series variance. Based on the Bai and Perron test results, we found two structural changes in tested returns mean series located in late 1998 and 1999. Additionally, the Inclan and Tiao Iterated Cumulative Sums of Squares (ICSS) algorithm test results identified some regime changes in futures crude palm oil variance located in late 1996, late 2001 and early 2008. We provide some internal and external explanations for the cause of these structural shifts in both mean and variance. Then, the study continues to investigate the implication of structural breaks in crude palm oil volatility clustering estimation process. Initially, we estimate a Baba, Engle, Kraft, and Kroner model (BEKK model) without the structural break. Having identified the breaks in the mean and variance of both returns series, we model their relationship by incorporating those breaks in the volatility clustering procedure (using modified BEKK model). The volatility clustering finding show that the spot crude palm oil persistency parameters have slightly increased when structural breaks are taken into account in the estimated model. In contrast to the spot crude palm oil findings, the futures crude palm oil exhibits a lower persistency estimation when structural changes are considered. The results support the importance of structural breaks in this volatility clustering estimation, and failure to do so may lead to bias persistency parameter estimation.