Predictability and Efficiency of the Philippine Stock Exchange Index

Jo-hui Chen, John Francis T. Diaz
(Chung Yuan Christian University, Chung-li, Taiwan)

Abstract: This study applied the autoregressive fractionally integrated moving average (ARFIMA) and the ARFIMA-fractional integrated general autoregressive conditional heteroskedasticity (ARFIMA-FIGARCH) models on the daily stock returns of the Philippine Stock Exchange Index (PSEi) from January 4, 2000 to December 29, 2011 and found significant long-memory processes in both return and volatility of the PSEi. The Chow breakpoint test was used to identify the structural break during the subprime mortgage crisis. Positive dependence in both return and volatility was evident before the structural break, but predictability was only observed in volatility after the breakpoint. The log-likelihood value consistently showed that the combined ARFIMA-FIGARCH model better characterized the PSEi.

Key words: long-memory; ARFIMA-FIGARCH; Philippine Stock Exchange Index

JEL codes: G11, G14