Application of AnyLogic simulation tool in the sphere of banking

This paper contains some banking issues that can be successfully resolved by the users of AnyLogicmultimethod simulation tool. The list of issues is not exhaustive, it reflects some tasks that have been discussed by XJ Technologies and AnyLogic users - banking organizations.


Current correspondent accounts balances form the essential part of the bank liquidity. High balances of the correspondent accounts lead to the yields reduction, but credit organizations tend to maintain high balances due to liquidity risks and ongoing payments support. We need to pay attention to the fact that the behavior of the items which affect the state of liquidity is quite stochastic. The absence of models which can forecast their behavior can lead to unnecessary interest expenses and insufficient yields. A simulation model analyses the current state of correspondent accounts balances, incoming and outgoing payments transactions, clients’ requests and forms a daily balance. This helps to maximize yields and guarantees a rated delay on the current payments.


It is inevitable that some stoppages during ATM network operations would appear. Stoppages occur when the ATM cassette is empty and the client receives a denial of the request for cash. The accumulation of the excessive amount of cash in the ATM is another unwanted situation. A simulation model can deal with a complex task: it can define a schedule of the ATM cash loading and proper volume of cash; it can also optimize the route of cash-in-transit guards who renew the ATM cassettes. The location of each ATM stations, the remaining amount of cash and frequency of usage are set at the input of the model. The definite schedule of ATM cassette renewal, the order of renewal and amount of cash for each ATM are formed at the output of the model.


Recently the new service of cash pooling has appeared in Russian banking system. Notional cash pooling has the company combine the balances of several accounts. This service allows holding companies to combine their subsidiaries’ accounts and effectively divide money between them to have a single overdraft limit. The part of money from master account is taken by the bank, which increases the bank’s yield. The simulation modeling task is to define the sum of money which the bank should earmark for the ongoing operations of the holding company. The quantity of current customers’ account balances, prospective incoming and outgoing payments to the subsidiaries’ accounts are set at the input of the model. The structure of the subsidiaries’ reserved money is formed at the output of the model.


Every bank tends to forecast the financial result and its fluctuations depending on some managerial decisions, such as the increase of a loan portfolio due to decrease of the loan interest rate. It is often made with the help of MS Excel and the software of statistic modeling. The model developed in AnyLogic simulation software provides much more accurate forecast of the financial result and helps to choose the most effective managerial decision. The model analyses the dynamics of aggregate financial results indicators within a definite period of time for the definite market behavior and the management of the assets and liabilities. The following market behavior scenarios can be used:

  • Normal state
  • Local crisis (cash flow)
  • Market crisis
  • System crisis

The detailed report on the properties of the Assets & Liabilities Statement for the definite time period (balance amount, average weighted interest rate)is formed at the output of the model.


The forecast of the bank financial performance for the quarter or a year and its comparison with the budget assignment for the corresponding period of time is one of the important issues. The following items can be used as financial indicators: Net Interest Margin, ROAE, ROA, Cost / Income, Deposit / Loans, Growth in Net Income. Estimation of the financial indicators can be delivered by simulation of the bank balance management at the required time period. The growth (reduction) scenario of the balance management statements (loans, deposits, current volatile customer accounts, etc.), changes of interest rates and nominal terms of financial instruments can be used as the model input parameters.


Risk-taking is the basis of banking sphere. The bank is successful when it adopts reasonable risks which dwell within its financial capacity. In order to implement this approach the integrated solution is needed, which performs data collection, forecasts aggregate risk indicators depending on time for various scenarios of the market behavior and asset-liability management. Simulation modeling and AnyLogic software form a good basis for creating a risk management model. The access to databases helps the modeler to process the detailed data on the level of accounts which give characteristics of the financial products. These characteristics serve as input data to the simulation of individual transactions. The indicators which correspond to the primary objectives of risk management, including the EAR (earning at risk), EVE (economical value of entity), VAR (value at risk), market assessments, indicators of the interest rate gap and a liquidity gap are formed at the output of the model.


Staff salary is one of the major items of expenses for any company. Therefore, large organizations with branch structure, such as banks which work with individuals, face the problem of business processes management. Determination of the tellers number needed for the effective work of the bank office is the example of such an issue. If the number of tellers is not sufficient, customers will not use the services of the office because of long queues and necessity to wait. If the bank tries to be overcautious and hire people abundantly, the bank will face the risk of inefficient business process with high level of expenses. A simulation model helps to analyze the number of people required at the definite time of day and make an effective timetable. There are various criteria of performance: the bank often tries to minimize the costs of maintaining the office by reducing the number of employees and stoppages during their work, controlling the average waiting time of customers. Scheduling is especially important in case that the same people may be involved in different processes. The examples of business process models can be found on the XJ Technologies website. Please open a Sales Funnel model - a model of sales procedure in the bank, which estimates the solvency of the client before issuing the consumer credit loan. The client can leave the office of the bank because of the unacceptable waiting time or the client can be refused to receive a needed amount of credit loan. The model allows us to estimate how the quantity and experience of the personnel can affect the results of the office performance.


IT resources management and IT services efficiency increase is a very important issue in the banking sphere. Frequently the IT infrastructure of a company is very discursively formed in response to some major business needs. As the result IT structure is quite a complex system from technical and economic point of view. There is a wide range of problems which should be advisably solved by simulation modeling, e.g.:

  • determination of the number of service-desk personnel required to ensure the efficient operation of the bank
  • estimation of the expected profit from investments in IT resources, such as hardware renewal
  • analysis of the effect of transferring some functions to external third parties (IT-outsourcing)
  • System crisis


One of the major simulation areas is the evaluation of the outcome from marketing activities implementation. Evaluation of profitability and risk of the new products launch (for example, a new loan proposal) or the effectiveness of advertising campaigns evaluation are the typical simulation objectives. Agent-based modeling is normally used to solve such problems. The model describes a typical behavior of a client (agent). The effect of the event implementation is evaluated on the basis of interacting agents.


One of the urgent issues in the banking sphere is the estimation of the cost (interest rate) of bank products with a high level of risk, for example, deposits with an early partial withdrawal or replenishment, credits which provide the possibility of early redemption. Additional features often lead to an increase in interest and liquidity risk, which must be compensated for by decrease of interest rates for deposits and increase of interest rates on loans. The calculation of the deviation of interest rates under different scenarios of market and customers behavior can be carried out using simulation, in particular, using agent-based approach.