Finity’s banking expertise
We support banks and lenders to better manage their risk, capital and improve profitability. Working in partnership with CFOs and CROs through to product managers and risk officers we provide access to cutting edge analytics to support revenue, capital and risk-based decisions.
Quantitative risk management:
We have developed quantitative models, including artificial intelligence to measure and manage credit risk. Our AI solutions have been deployed across underwriting, fraud and anomaly detection.
Capital management:
We apply a business lens to capital management to support risk appetite, business and strategic objectives. Our expertise covers stress testing, internal ratings based capital, economic capital and loan loss provisioning.
Advanced analytics:
We support banking clients to develop algorithms to optimise decision making and automate tasks in domains such as loan pricing, responsible lending, and customer retention.
Software development:
We can assist with software development and have experience building tools such as production-ready APIs to implement artificial intelligence solutions.
Experience in banking
Credit risk modelling
We have developed credit risk models for a range of lenders including:
- Internal Ratings Based models: Developing Probability of Default, Loss Given Default and Exposure at Default models that underpin Internal Ratings Based capital calculations.
- Credit Scoring Models: For use in loan scoring and pricing. We have expertise in both traditional statistical models as well as the application of novel artificial intelligence models utilising alternative data sources.
- We understand regulatory requirements for banks to manage model risk and work collaboratively during our engagements to ensure that models meet the bank’s model governance requirements.
Bank capital modelling
Our expertise covers capital modelling and management.
- We have developed and validated models for capital calculation, stress testing, and economic capital for banks.
- We have worked with banks and LMI providers to develop economic models that translate movements in economic variables such as GDP, unemployment, interest rates and house price index to financial metrics for banks such as arrears, credit defaults, and net interest margins.
- We have also developed bank wide models to forecast the profit and loss and balance sheet positions to understand the capital position of the bank under each scenario.
Advanced analytics
Finity has extensive experience in customer and pricing analytics and use applied machine learning techniques to help lenders grow their portfolio and improve profitability. We have developed datasets using granular demographic and sociographic data which can be used to complement a client’s internal customer data. Our banking experience includes applying machine learning to model and predict:
- Loan default probabilities using a range of borrower related data
- Forecast behavioural characteristics such as prepayment and early repayment to build detailed profit and loss metrics for each potential borrower
- Demand elasticity models to determine optimal borrower rates for each client. Our optimisation approach also incorporates the lender’s organisational and platform constraints in recommending the optimal price to offer each potential borrower.
- Predict the risk of customers exiting at the end of their loan term and then suggest individualised actions to improve retention.
Australia’s analytics pioneers
Our pricing analytics experts pioneered the use of machine learning algorithms to optimise pricing for financial products. We have invested in developing proprietary data assets such as socio demographic datasets, geospatial climate risk models and market price monitoring software.
Finity purchased AI firm Deep Logic in 2019, expanding our capabilities in this area, giving Finity access to leading expertise in artificial intelligence and machine learning, data management and software development.
Case Studies
Pricing analytics for retail loans
Project
We worked with a retail lender where we used machine learning to set risk based prices. The risk based pricing model used machine learning to calculate a default probability given a range of customer related data.
Optimisation approach
We forecast behavioural characteristics such as prepayment and early repayment to create a profit and loss statement for each client. A risk based pricing model was combined with demand elasticity models to determine optimal borrower rates. This approach incorporated organisational and platform constraints on prices. The model leveraged our proprietary household level socio-demographic database, Defin’d in conjunction with customer level data collected by the auto loan provider.
Result
The methodology allowed the lender to increase volumes while maintaining profitability in a relatively competitive marketplace.
Improving Customer retention
Project
We have developed a retention model for retail lending clients that applies machine learning to understand customer behaviour and suggest actions to improve retention.
Approach
Our approach identified characteristics and factors of individual customers that were predictive of whether a client continued with other products with the lender. Once the model was constructed, Finity worked with the client’s platform vendor to embed the predictive models.
Result
Our model allowed the lender to identify the next best offer for clients approaching the end of their loan term and to prioritise customer activities. The model supported the client coordinate its retention activities across multiple channels.
Talk to us about Banking
Sen Nagarajan
Mobile +61 434 195 740
sen.nagarajan@finity.com.au