The economic impact of the pandemic has affected mid-sized businesses in many different ways. Greg Beamish, Head of Credit at ThinCats explains how his team responded to the pandemic and continued to provide funding support in such a fast changing and uncertain trading environment.
How can lenders use data to improve service?
Lenders having access to important borrower data can improve the service they provide in numerous ways, from quicker credit decision making to reducing the ongoing administrative burden for their clients. We look at some of the improvements we made in 2020 and plans for the future.
Our assessment of new funding proposals combines the expertise of our business development and credit teams in understanding the business models of our borrowers with our big data-driven machine learning tools to determine the overall credit risk. Our sophisticated data models are collectively known as PRISM and we have made some significant enhancements during 2020.
In June we launched a new version of our proprietary credit model PRISM Risk Grade which we have proven to be a very accurate predictor of defaults amongst midsized businesses. The predictive accuracy of the new model was tested over the last 12 years and was shown to retain its accuracy throughout all stages of the economic cycle, across all industry sectors and sizes of businesses within the mid-sized SME universe which we define as companies with £0.5m-£40m in total assets.
Furthermore, the model is predictive not just of defaults over a 12-month window from point of calculation, but also over longer time frames (up to four years), so is highly suitable for assessing credit risk over the typical term of loan that we offer our borrowers.
During the year we enhanced our Pricing Portal which is the tool that calculates the price of a loan for a specific borrower based on a combination of PRISM Risk Grade and Security Grade data. New functionality has been added to enable more recent or more complete financial information to be captured than is available through filed financial accounts. Further information relating to the purpose of the loan is also captured enabling us to estimate the impact of the proposed loan on the borrower’s balance sheet.
We have also created a new credit dashboard within the Pricing Portal to enhance the underwriting team’s understanding of how each risk grade has been calculated and allow them to run what-if analyses to test alternative scenarios. The dashboard also provides benchmark data comparing the borrower to other firms in its sector in terms of key financial KPIs.
In the autumn we trialled and rolled out Open Banking as a condition precedent for all new loans. For a lender such as ThinCats, which funds a large number of cashflow backed loans, having access to the movements of a borrower’s current account provides valuable data to support more informed credit decisions and to set covenants that are better aligned with the borrower’s business cycle. The platform also reduces the ongoing administrative burden for borrowers in providing regular MI post drawdown.
Although much of our data analytics activity happens “underneath the bonnet”, the common thread across all these enhancements is pulling together and interpreting multiple data sources so that we can make quicker and better-informed lending decisions specifically for mid-sized businesses and their advisers as well as structure transactions that better align with the company’s needs. As we implement further developments through 2021 we plan to share our data insights more widely with the business finance community.