Just as sports franchises have learned to use data analysis to assign values to skills that had been difficult to judge previously, financial institutions now can discern the most and least productive aspects of their loan portfolio, and act accordingly.
Thanks to the wonders of GPS, you can ascertain how much ground a major league baseball outfielder covers to catch fly outs, the exact spot on the floor where an NBA player posts his best (or worst) shooting percentage, and how fast a football player streaks down the field on a kickoff unit, or how consistently he runs routes.
In and of itself, the information is interesting but not actionable. The value emerges when you compare those values against other players’ performance — and weigh certain skills against other skills to see which has the most impact on whether you win games.
Likewise, thanks to conjoint analysis, credit unions possess the capability to make well-informed judgments about which segments of their loan portfolio provide the best return, bringing clarity to what historically would’ve been difficult comparisons. Conjoint analysis is a statistical technique that determines which combination of attributes in a defined data set is most influential.
In the world of lending portfolio management, this technique plays out as static pool analysis — the performance assessment of a specific group of similar loans over a specific period of time, such as home equity lines of credit from 2010 compared with those from 2012 or 2014. The more precise the pool, the more precise your estimates will be.
Static pool analysis serves several purposes: to calculate and manage risk, identify existing profitability factors (fees, loan-to-value assessments, risk-based pricing), recognize potential profit centers, predict and potentially prevent future losses, and provide regulators documentation of sound loan portfolio management processes.
Reports can be categorized two ways: single pool analysis or pool comparisons.
Single pool analysis reports can be presented as a point-in-time report that assesses data such as concentrations, delinquency, profitability, and the like. Or it can be presented as a trending report, which demonstrates how the loan pool has fared on key metrics from the cutoff date to the present day. These reports offer the capability to forecast patterns such as loan repayment speed.
Meanwhile, pool comparison reports offer a side-by-side look at loan pools at a point in time or a similar timespan after the cutoff comparison. Both approaches have validity — the former as a checkpoint for current portfolio health, the latter as a measure of performance at the same level of maturity.
Two key points on forming these loan pools: You must adhere strictly to the start and stop dates, for sake of consistency; and the internal factors (for instance, underwriting changes) and external factors (such as interest rates) affecting various pools should be similar or at least taken into consideration.