Behavior Investigation of Islamic Bank Deposit Return Utilizing Artificial Neural Networks Model

Abstract

According to Islamic law, the Islamic banks deliver return to depositors based on profit and loss sharing principle. Consequently, the return will be uncertain following real business and economic condition. However, this research investigates the phenomenon that Indonesian Islamic banking industry seems to mimic interest rate in generating return to depositors. This investigation utilizes artificial neural networks (ANN) model to examine the importance rate of each macroeconomic variables used. The rate indicates the level of domination or contribution of each variable in determining the volatility of Islamic time deposit return. The research uses ten years of monthly macroeconomic data as independent variables. Additionally, the average rate of return from one-month time deposit of all Indonesian Islamic banks (RR) is used as dependent variable. As a result, the N^(9-4-1), a chosen neural networks architecture, found that BIRT and INTR as proxy of interest rate, dominantly affect RR volatility with almost 98% of importance rate. It shows the very high dominance of interest rate in determining RR volatility, as indication of mimicking behavior, rather than remaining variables as proxy of real economic condition.