Prediction of Optimal Cash Injection in Automated Teller Machines, ARMA Approach

The development of electronic banking services, in order to facilitate the monetary transaction viz-a-viz banking system, requires the scientific and efficient management. The Automated Teller Machines (ATMs) which are mostly deployed in various banks branches, and are subject to direct supervision of branches authorities, will not get cash optimally injected. This mode of liquidity management, of ATMs will not only distort the brand name of banking system in a psedue competitive market, but it may lead to inefficient management of liquidity through the cash injection more than the daily need of these machines. Hence in this paper an attempt is made to predict the optimal injection of cash to ATMs of Sepah Bank by using the daily time series data of withdrawing money from them. For this purpose, we applied two econometric methods of ARMA modeling techniques. The results prima facie indicate that, despite existence of certain constraints on statistical data due to non operation of ATMs, impairment of Shetab Network, and delayed feeding of cash, the lower range of prediction error, is about 1%. Thus, we conclude that, to optimize the cash injection process of ATMs, besides considering the human resources and required security measures, the application of above study is recommended.