VARS – VakifBank Risk Solution Qorus Banking Innovation Awards 2023

Banks are required to calculate expected credit losses for their financial assets within the framework of specific accounting standards in order to account for potential risks that may arise in the future. Specific statistical methods and techniques are determined when calculating expected losses. Calculation applications are needed to implement these methods. Our bank used an external application to calculate expected credit losses for future potential risks. In that tool:  It was not easy to implement and deploy model changes.  Implementing new model was open to make faults.  When the application received an error, it was not easy to find the cause of the error.  Local support was not adequate. We had to tackle problems with ourselves.  Limited number (Only 2) of users could be able to create and run their scenarios. If more than 2 users wanted to run their scenarios, the had to wait completion of others.  It was not possible to run the models on a daily basis because it took too much time to complete them. When the execution failed, all the cycle had to start from the beginning.  Expected Credit Loss could not be run only for specific customers. Even if you want to calculate only 1 customer, you have to run all customer. To do this some code changes were needed. For above reasons we started an initiative to build our own solution named VARS. VARS is a tool to calculate Expected Credit Loss and it is developed to help the bank comply IFRS 9 requirements. VARS was also developed in house. It allows implementing complex models. It also allows business user letting them to get results quickly by using its computational power. Thanks to VARS for IFRS 9  Large number of complex models can be implemented, executed and applying accounting rules easily.  Model implementation and execution risk reduced.  Expected credit loss are calculating on daily bases. Business users are able to see expected credit loss on daily basis.  Business users are able to create their scenarios themselves without it support.  Business users are able to calculate expected credit loss only for specific risks if needed without it support.  Business users are able to experiment new scenarios by changing some parameters themselves without it support.  Model change requests are implementing in reasonable time and effort when they occur.  Business users are able to access risk reports easily.  Business users are able to follow workflow.  There is no need to restart the entire job when execution fails. The job can continue to run from the point where the error occurred.