Challenge

The ALM (asset liability management) process can not be considered easy. Generically speaking, ALM goal is to allow banks and companies Treasury to take control over liquidity risks. Also, being a managerial control, it allows some inferences about regular operations and some multiples.

To make more information available from his ALM, Toyota Financial Services decided to create an ALM simulation system. This system should be able to forecast earnings and others useful indicators as well as it should have some conditions to behave as close as possible to the reality. Settings should be placed for others Bank sectors, like sales forecasts or even exchange rate scenarios.

Also, a stochastic model for interest rate with not 1 or 2 scenarios, but 5,000 or 10,000 would be a plus.

For that, Toyota Financial Services contacted Élin Duxus: to create a robust, reliable and safe ALM Simulation.

How Élin Duxus Help

Élin Duxus and Toyota Financial Services Treasury team developed a methodology that became into a system capable of run a simulation for Bank ALM.

This system runs till 10,000 stochastic scenarios for interest rates (and the balance at all) and reach 10 years ahead, having settings for all assets and liabilities, including exchange rates, loans, portfolio of clients, swap and equity.

It is possible to run a simulation from an initial date and see till the PDF (probability density function) of the results and also the distribution for some important indicators.

Another interesting feature is the system speed. The application was concept in C++, a very low level language, what means, a really fast and professional language. And the system can run in a multi-thread way, without blocking others applications it shares.

Finally, to allow all settings and result and to interact with others sector of the Bank, the system uses the same database that Bank uses, meaning less efforts for Toyota Financial Services IT area.

Value Added

Some may say that ALM Simulation is not complete, but, in fact, is a great system. The total system amount of settings and possibilities – despite the fact that a user must be trained to use it – allows a very precise copy of diary Bank operations. We must notice that this is a simulation system and depending on the settings, results must be throw away. In other words, garbage in garbage out. But, realistic settings have shown such a confidence in this application that, as a consequence, studies are taking place to spread the system beyond Brazilian limits.

This flexible system also generates some critics about regular operations. These critics added with some pretty nice ideas that come to life during developing and during system utilization are, together, creating a very clear view of real ALM traps and making ALM a very safe and controlled corporate tool.

This system spent about 2 years to be studied, developed and fully implemented by Toyota Financial Services and Élin Duxus.