2. Data analytics
If it hasn’t already, your credit union should take a serious look at soft ware that analyzes the huge amount of data that exists across numerous soft ware applications. This data—properly analyzed—can be a tremendous tool to help you make decisions and develop strategies. And the need for more aggressive risk management solutions requires greater analytical tools to help manage and mitigate risk.
Credit unions have extensive data on, for example, loan performance and members. Unfortunately, this data oft en is housed across multiple soft ware applications that rarely communicate with each other.
Forward-looking credit unions realize the strategic competitive advantage that’s obtained by using a single platform to aggregate this data and apply reporting and analytical tools at the loan and member levels. Users from all departments benefit as they gain new perspectives of their portfolios and other relationships.
This is generally called “data mining” or “data analytics.” Now, with accessible loan-level and member-level data, managers of all departments can identify risk in greater detail. They can also discover untapped markets, product and pricing opportunities, and revealing statistical trends. This level of reporting detail also helps satisfy compliance requirements.
Many credit unions use historic performance methodologies to estimate reserves and manage loan loss allowances. But using a predictive model with all available data can give you a more complete picture of a loan and its expected performance.
Credit unions can use data characteristics from borrowers, loans, collateral, regional geography, and the economy to develop a probability of default and loss severity for each loan. You can apply these measures to future cash flows (performance) at the loan level based on potential economic scenarios to estimate loss and your loan loss allowance.
The success of predictive analytics depends on the vast array of centralized data you have. If you have robust data and advanced analytical tools, the accuracy of your predictions will be high.
For the borrower category, some of the more common data points available for a predictive model are income, debt, direct deposit, delinquencies, credit score, and employment. For the loan category, you can include balance, rate, term, pay-date trend, auto pay, superior/subordinate loans, and length of time delinquent.
Examples of collateral data points include value, liquidity, value trend, and marketing time. A predictive model also includes institutional or credit union attributes to profile members, such as borrower stability, charge-offs, borrower bankruptcy rate, and delinquency rates.
Predictive methods also can be applied to lending. Today, the fastest loan decision made with appropriate risk-based pricing oft en determines who originates the loan. Predictive modeling can provide operational advantages to enhance not only speed but the confidence level at which a suitable risk-based price is offered.
Using the appropriate system to aggregate data into a single relational database provides a competitive advantage to credit unions that successfully use data mining and analytics to turn knowledge into strategic decision-making.
3. Loan participations
The challenges of lending in today’s economy combined with weak investment returns are prompting CFOs to get more involved in loan participations. These instruments can add both diversity and income to your loan portfolio.
Government-guaranteed loan participations oft en call for a CFO’s involvement. These instruments don’t count against your regulatory business loan cap, and they have significantly less credit risk because they’re guaranteed.
Although the returns are lower compared to most participation loans, these instruments provide much better returns than comparable investments. There are three types of government-guaranteed loan participations:
1. U.S. Department of Agriculture (USDA) Business and Industry;
2. USDA Farm Ownership; and
3. Small Business Administration (SBA) loans.
The investor purchases the guaranteed portion of the SBA/USDA loan from the originating lender and receives an unconditional guarantee: no credit risk so long as the investor isn’t knowingly participating in a fraudulent situation. And the unconditional guarantee from the government transforms the loan into a permissible investment under NCUA rules.
Unlike traditional participation loans that require full underwriting, government-guaranteed participations require no credit underwriting review. Evaluating these loans is much like appraising a security based on a yield table rather than examining a borrower’s credit profile.
Government-guaranteed participations can be considered an investment or a loan, but a loan not subject to traditional participation and lending regulations, according to NCUA. In addition, there are active markets to sell these participations if the need for liquidity arises, unlike traditional participation lending.
Most SBA loans allow quarterly rate adjustments with no periodic caps and floors. And USDA loans are fixed for a short window, then adjustable thereafter, which certainly helps in interest-rate and price sensitivity analyses.
Government-guaranteed participations are more like investment securities. This lets CFOs use their ability to analyze characteristics, such as prepayment assumptions, and to amortize any premiums paid over the average life of the investment.
As in any investment, diversification over multiple guaranteed loans is prudent. There are penalties in many loans that can stem early prepays and, therefore, reduce accelerated premium write downs.
NEXT: Surplus funds