Changing Lending Strategy: Static Pool Delinquency Measures Can Help You

By Michael Cochrum, CUBI.Pro

Recently, I’ve been speaking to credit unions who have announced a major shift in lending strategy.  After many years pursuing auto loans in a highly competitive indirect market, several of my clients have decided to take a breather and concentrate more heavily on lending directly to their members.  This shift comes because of several factors, including a slowing of vehicle sales shrinking the marketplace and increasing competition for loans, rising rates which have temporarily shrunk yields as portfolios adjust to rising costs of funds, and in some cases, fatigue from constantly battling competitive forces over several years.  While all these reasons provide justification for change, it is important to understand the effect that a rapid change in strategy will have on portfolio performance in the short-term.
While a rapidly growing portfolio can hide, or dilute, delinquency as new loans increase the denominator used to calculate a delinquency ratio, the opposite can happen as portfolios begin to level-off or begin to shrink and delinquency becomes more concentrated.

The chart above demonstrates the effect that lagging delinquency statistics can have on the perception of portfolio performance when measured using a simple delinquency metric over time.  We could leave this conversation here and agree that we all understand this, intuitively, except that individuals have actually lost their jobs because they couldn’t explain this behavior, nor could they demonstrate why board members and regulators shouldn’t be concerned with “rising delinquency”.

Let’s begin by reminding ourselves of what it is that delinquency is used to measure.  It is supposed to measure the percentage of loans from an origination pool that did not make a payment by the due date.  Reportable delinquency, for credit unions, is the amount of loans that are sixty or more days past due, compared to the total amount of loans in the portfolio.  Delinquency is used as a “proxy” to predict future loss.  In other words, most all loans that charge-off were first delinquent.  Therefore, as delinquency increases, the probability of losses increasing is higher - except, of course, when it isn’t. 

When portfolios are increasing rapidly, as when a new program is started, or decreasing rapidly, as when a program is halted, the predictive value of delinquency becomes skewed.  The way around this phenomenon is by using a different metric.  In this case, I would recommend using a static pool delinquency metric, versus the traditional delinquency measurement.  In this case, you would be measuring delinquency against the balances of a pool of loans originated during the same time-period, rather than including recently originated loans. Thereby, eliminating fluctuation based on the increase and decrease of originated loans.  Further, a static pool metric allows one to more accurately determine the performance of loans during the same time-period, unmasking the impact of policy and strategy changes on performance.

In the example above, one can see that, regardless of changing originations, each pool of loans has roughly the same delinquency performance over time, except for the 2014 origination pool, which has experienced much higher delinquencies than the other pools observed.  Using a static pool metric is much more valuable in assessing true delinquency trends than simply calculating current delinquency each month against a fluctuating portfolio balance.

If you have questions about using static pool metrics, rather than your traditional metrics, please visit us or call 866.422.8720.

About the Author:

Michael Cochrum is CEO/CDO of CUBI.Pro, a company that specializing in working with financial institutions on data transformation projects.  Michael has worked in the financial services industry for almost 30 years and with, or in, credit unions for the last 20 years.  He holds a B.S. in Data Analytics and earned his MBA from Texas A&M – Corpus Christi.  You can email Mr. Cochrum with questions at or visit the website.


Popular posts from this blog

How to Estimate the Impact of CECL on ALLL

Measurements Matter

Are You a Coach or a Referee?