There are many important aspects for Consumer Lending Analytics. Here three of them were touched. The first being ‘Income Analytics’, second ‘Journey of Applied Analytics’ and lastly, ‘Analytics Driving Business Decisions’.
The first aspect which was deriving the value of analytics around income mainly concerns unsolicited credit card line increase program. The core area of concern here is that without causing inconvenience to members and regulatory problems along with reduced credit risks how can we know the current income of borrowers. The solution here is assessing the already existing data with us. This covers four types of incomes which are- verified income which consists of mortgage system, stated income which is all income sources over the last 12 months, validated income mainly comprising external deposits, recurring, payroll, social security, and direct deposit and lastly the income data from the credit bureaus.
The main task here is of generating value through maximizing experiences of existing members. Collaboration and purposes are critical to the success of many things- solving a business problem, achieving a positive rate of interest, enabling enhanced member experience and mitigating regulatory risks. The two perspectives to be considered are interchanging perspective and income perspective. By solving the income problem, the credit limit can thereby be increased and this, in turn, results in increased spending of our members ultimately improving our profitability. The advantage of income analysis is that the credit unions have their data, so there is no need to deal with regulatory requirements in addition to using this income data for other purposes.
Income Analytics can not only be used in a Credit Line Increase Program but also in promotions and campaigns, pre-approvals, automated loan decisions, portfolio segmentation, portfolio risk management, and proactive member assistance programs. All in all, the path to CLI Program starts and ends with data analytics. There is a series to be followed- beginning with Income Analysis followed by Credit and DTI Analysis, Technical Specifications Documentation, Program Documentation, Member Communication Plan, Compliance Review, Core System, Increase Script, Staff Readiness, CLI Executed, Measuring Feedback and accomplished with Analyzing Portfolio Performance. The result of these efforts being strong financial returns.
The second aspect relates to ‘Transactional Analytics.’ The current state of Lending Analytics is providing valuable insights into the impact of lending decisions on loan performance and significant latency between insights and action. Progress has been from reports and dashboards to analytics workbench. The important questions here are how to do a better job at automating approvals in the lending decision without necessarily increasing risks and how to re-calibrate the underwriting decision criteria that goes into making that automated decision. The solution is to drive more automated approvals without increasing risk. The value of near real-time analytics is leveraging machine learning and decision automation to provide advanced analytics and reducing latency from insight to action.
The key at this juncture is monitoring information portal, exploring analytics workbench, investigating data science labs incorporating advanced analytics and machine learning and industrializing decision hub. What is needed to increase the value or performance is a primarily copious amount of data for performance detection which includes host data relating to loan performance data and LOS data to correlate more granular loan origination data with loan performance data. Data diversification or in simple words lots of data which is aggregated across credit union is required to get insights. Secondarily machine learning capabilities are necessary to determine the best fit and to accelerate the correlation of data into more meaningful insights and then moving on to the decision hub. Decision Hub uses Data Science Laboratory with Machine Learning so that results can be delivered through APIs directly to loan origination systems and decision engines. It drives out the latency between analytic insights and operationalizing the insights. The benefits of adopting this are more loans with less risk, increased operating efficiency and much better member experience.
The last aspect is ‘Using Analytics to Drive Intelligent Loan Portfolio Management Decisions’. The issue with this is what we are doing with Business Intelligence in Credit Union space. At this point, we need to distinguish between User standpoint which existed nearly 30 years ago and the Technology standpoint. The situation with CU risk-based pricing or lending is that strategic decisions are being made about lending without full exposure and understanding of relevant data. After loan origination, very little risk evaluation is done and underwriters are often excluded in the feedback loop. Pricing does not consider all risk and these impacts loans. We need to shift from risk-based lending to risk-based pricing because according to risk-based pricing yield at the end is supposed to be somewhat equivalent and adjustment of rate is supposed to be the risk and not for profit.
There are certain elements to be thoughtful of while formulating a strategy. One is Risk Modeling to know the real life cycle risk at the time of origination to ensure accurate pricing. Second is Automated Decisions to reduce the error rate caused by reliance on human intuition and increase the time spent on marginal credit decisions. The third is Risk Review to update significant characteristics and monitor risk based on current risk profile. Fourth is to provide Underwriter Feedback to provide usable information to underwriters to make educated credit decisions and last to Renew Strategy to mold future strategies based on information from past experiences. There should be a partnership between Credit Union and the business side of the house. People with experience should be used as intermediaries for interpretation between the technology and the business use of that technology.
In conclusion, algorithms or computer programs should be used to drive decision making utilizing people, process and technology. New challenges are arising from risk lending, therefore, the need to do the job more efficiently. And if pricing is a competitive advantage, then the costs should be lowered by replacing human beings with computers. Nowadays, clients have the technology but not fully implemented, and credit unions step in to do that. Most important is closing the feedback loop for educated decisions.