Artificial Intelligence for Islamic Sukuk Rating Predictions

Abstract

The development of a Sukuk market as the alternative to the existing conventional bond market has given rise to an issue of the availability and accuracy of some of Sukuk issuance’s ratings. This results in asymmetric information among capital market players, which ultimately needs to be reduced. Moreover, the Basel III framework allows banks to establish capital adequacy requirements based on ratings provided by external credit rating agencies or, in the absence of these, to determine ratings of its investment internally. For these reasons, ratings are considered important by issuers, investors and regulators alike. Focusing on Malaysian outstanding long-term corporate Sukuk in the year 2012, this study tests the efficacy and accuracy of the Sukuk rating model when compared to the ratings assigned using a Multinomial Logistic Regression model, a Decision Tree and a Neural Network. In order to address the limited study on Sukuk rating prediction, this research provides an empirical foundation for the investors to estimate the ratings assigned. The study examines variables from past research on rating prediction models taking into account various Sukuk structures, credit enhancement facilities, the industrial sector and macroeconomic variables. All methods strongly indicate that share price, Sukuk structure and guarantee status are empirically proven to be key factors when predicting Sukuk ratings. Furthermore, neural network methods obtain the highest accuracy rate when predicting the actual rating in the market compared to the other two methods. Therefore, it is expected that the proposed models are beneficial to the rating agencies, Sukuk issuer companies, corporate managers, private and institutional investor to support their investment decision making. The regulatory agencies may also take advantage of these models as they can be used as a benchmark for the Internal Rating Based) approach as required in Basel III. In line with those practical implications, this study also aims to contribute to the development of an Islamic finance body of knowledge.