A Shariah compliant alter ego for the Dow Jones Industrial Average

Dow Jones industrial average is a reputable 30-stocks investment vehicle constructed through a multi-criterion “human expert judgement”. However, more than half of its constituents are non- compliant to Shariah standards and a na ̈ıve Shariah compliance filter of DJIA may render a less diversified portfolio. We develop a Shariah compliant alter ego for Dow Jones industrial average (DJIA) that mimics its stock selection process. We define the alter ego as a linear system that selects top ranked securities for their performance in a set of well defined features including Shariah standards. The features and their weights in the system are defined separately using supervised machine learning models against DJIA committee decisions. The features are extracted once dur- ing the study period but their weights are defined at each decision point by a linear discriminant model trained for an expanding window of previous available data. The system selects 30 Shariah compliant reputable securities from S&P500 at each DJIA reconstitution date that may be named as Shariah compliant industrial average (SCIA). The SCIA remains as similar as possible to DJIA in terms of the features, sectoral diversification and portfolio performance