In this session, panelists- Dave Ulrich, Karan Bhalla, and Matt Maguire discuss analytics and the various credit union segments where they are being used. These three practitioners talk about three particular use cases as discussed below.
Credit Risk Analysis and Preparation for CECL by Dave Ulrich, Head of Credit Administration at First Tech Federal Credit Union,
Why CECL?
CECL stands for Current Expected Credit Loss, and it's a new accounting standard that will change how financial institutions account for expected credit losses. It replaces the current standards for loss accounting – commonly known as FAS-5 and FAS-114. Dave suggests that through an incurred loss model, we figure out what kinds of parts of the portfolio have already experienced some sort of loss and estimate what the impact for that is going to be. CECL forces us to confront the question, “ would you buy your own loan portfolio?”
Answering this question will lead to an increase in our allowance because an incurred loss is a subset of a lifetime loss injury.
When the credit cycle peaks, our reserves drop and so when the cycle turns, the funds are not there, and it continues to be a vicious cycle.
One way to mitigate is CECL, which implies that we need to increase our reserves.
By displaying multiplier against current allowance mode for standard portfolio loan to estimate an increase in the loss reserve and shows that when CECL gets implemented there is an impact on the order of significant expenses. Where are we on the current credit cycle? At the peak, drop or middle is the question that needs to be answered by observing the multiple time frames.
What first tech is doing about it? Expected outcomes and benefits.
For first tech, Dave explained how through a model inventory he was able to analyze the credit products divided into four categories - Mortgage, HELOC, Auto, and Card, ( which represented 90% of their loan portfolio) , and by the simple red yellow green ranking, he was able to claim that auto portfolio was the most robust for First Tech but they face a huge issue on the mortgage side, where they believe that CECL is also going to have an impact. A huge amount of mortgage leads to a negative impact on the balance sheet.
First tech is good with data and loan level transaction data can be accessed easily but the issue that lies is that it is not a full credit cycle. The advantage of working with deep future analytics is that they are building a consortium of data and then building models on top and we end up contributing to that database.
The four c’s- collect, cleanse, calculate and consume, this is the lifecycle they are going through with their data.
With the implementation of CECL, it is possible to allocate capital for potential losses. Loan-level pricing can be leveraged in pricing and stress testing.
Loan-level models are essential, not just in terms of building out robust CECL forecast, but once you have got a handle on the cash flows all the way down to a loan level, that allows you to do a lot with the information and you can support capital planning and enhance risk-based pricing. Once you get down to the loan level, the world is your oyster.
Credit unions can use analytics to win by Karan Bhalla, CEO, CU Rise analytics
Debit Card programs and Credit Card programs and how analytics can drive specific value through that particular channel and use.
Some Barriers faced by credit unions when it comes to beginning these initiatives-
- Where to start?
- Organizational buy-in
- Inadequate resources and technology
- Easy marketing.
Case study 1 - Debit rewards campaign
- CBC had lower than desired debit card penetration among its members.
- Members with debit card contributed sizeably to interchange income.
- Some CUs and banks have started offering debit usage incentives, No CU in Ventura country had any such offering.
- To increase the usage of debit cards and gain a competitive advantage, it was discussed to launch debit rewards programme in CBC.
Approach
- Data collection - Analyzed debit users and their activation rate. Calculated interchange income from these debit users.
- Research - Identified other credit unions offering incentives on debit usage.
- Product development - Identified various offers for debit rewards
- ROI calculation - Calculated reward cost for each of the offer identified. Recommended Nickel back for both signature and PIN-based transactions.
Case study 2 - Credit Limit Increase.
- To keep CBC’s credit card top of the wallet it has to be competitive enough when compared to other options available to the member.
- In order to serve this purpose, regular credit limit increase campaigns are carried out for deserving card holders.
Approach
- Data collection - Analyzed performance history and bureau history of cardholders.Excluded card holders depicting risky behavior.
- Analytics assessment - Using decision tree technique, portfolio divided into meaningful small segments, so as to decide a strategy for a limit increase.
- Recommendations- Based on strategy decided and cardholders existing credit limit, new credit limits assigned.
He discussed the ongoing analytics projects at CBC and their results, which included a 25%+ increase in assets and members and net worth up by 34%.
This is how the journey should start -
- Commit to a data-driven culture.
- Start by allocating a small analytics budget.
- Obtain the right talent.
- Take on manageable projects.
- Monitor, review returns and learn.
4 Ways to make Data your Best Digital Transformation Asset by Matt Maguire, Chief Data Officer at CO-OP Financial Services
Top of wallet becomes top of the device. Example - American Airlines and their app revolution, Amazon, Netflix, Spotify and how they are using data for predictive analytics. They are focused on their space and are suggestive for their customer.
“The magic of the data is the ability to predict the future.”
Here is how Matt suggests we get working on this magic.
1. Establish a Data Strategy Based on Solving a Problem
The best strategies are specific and measurable. Member’s problems should drive the solutions. The innovation needs to be relevant.
2. Separate the signal from the noise.
Find actionable intelligence in the data. Making randomly available data relevant is the key.
3. Get your data governance locked down early
The four key priorities should be - end to end traceability, quality, and controls, privacy and security, accessibility.
4. Build and foster an analytics culture.
The four key attributes towards building and fostering such a culture are - Collaboration, entire cooperative shares in the asset, enterprise-wide understanding of the goals, and a shared vision.
5. Integration is Innovation
Dig deeper and analyze the service you are providing and what more value can you add. Evolve and innovate to create a seamless experience.
The panelists provide extraordinary solutions to problems by involving data and showcasing how data can have an impact differently, on these different areas. They display how the field of data analytics is continuously evolving and the technology and science component is constantly moving.
Comments