The performance of BB-MCEWMA model: Case study on sukuk Rantau Abang Capital Berhad, Malaysia

Abstract

Monitoring process of auto-correlated data has become an impressive impact in Statistical Process Control (SPC). It is due to the nature sensitivity where the current information data is said to be correlate to previous data, and well as the data can be categorised as dependent data. Thus, this study considered an alternative chart and particularly known as MCEWMA chart. The function is to quick detection on small shifts in the monitoring process. However, monitoring the autocorrelation having an indirect problem where it is often influenced by inaccurate estimation from its base model. Ignoring this problem will eventually cause another problem such as low performance of model where it said to be less effective. Thus, to solve this problem, a double bootstrap approach is hybrid into the MCEWMA chart. The main objective of this study is to construct a standard error of confidence intervals of MCEWMA. The algorithm in this study eventually comes out with a new model coins as double bootstrap MCEWMA (BB-MCEWMA). To see the effectiveness the proposed model, real data will implemented, which sukuk (Islamic financial instrument) is issued by Rantau Abang Capital Berhad, Malaysia. The trade date for the sukuk is taken from March 2006 until March 2011. The performance of BB-MCEWMA in terms of effectiveness of point estimator and interval estimator (Normal, Student’s-t and BCa) for the proposed model and the MCEWMA are considered. BB-MCEWMA gives smallest value of error (MSE and RMSE) and shortest length of interval estimation. With these result, it is proven that the hybrid of double bootstrap approach into MCEWMA model fixed the estimation of original model and single bootstrap approach model, and statistically give more accurate estimation model