Credit unions may be faced with implications that last longer than the typical economic recession.
Having 40 years of lending experience, including 35 years of C-level experience overseeing lending at two different credit unions, I can safely say I’ve never seen a set of economic conditions like what we’re experiencing now. Whether it’s the rapid rise in interest rates instigated by the Fed, a widening of the gap between the haves and have-nots (perhaps partially explained by this post I made last month on LinkedIn) along with the lack of affordable housing, this is all new to me, as it is to you.
Yet the most perplexing of the economic conditions I’m witnessing now is the rise in consumer credit losses. I’ve never seen losses rise so quickly in such a short period when the country is not experiencing a recession and job losses. About six months ago, I jokingly started calling what Ent Credit Union and all lenders are experiencing a “consumer credit recession.” Except now, I’m not joking: This decline in consumer credit performance is unprecedented. Here’s my best diagnosis of what’s driving this degradation of consumer credit performance.
Is the highest inflation this country has experienced in more than 40 years partially driving a consumer credit recession? As discussed in this article from CNBC, the card industry is trying to diagnose why losses are increasing at a time the economy is still generally healthy. While the article seems to indicate that the rise in losses is due to the rapid rise in balances, if we look at credit card debt levels prior to COVID-19 in this article from CNN, it appears balances have increased about 11% in 3.5 years. This is not quite so dire a situation as it might seem because personal incomes have risen even faster during the same period.
Due to the previously mentioned benefit homeowners are experiencing if they refinanced during the COVID-19 years, along with the likelihood that inflation in rent, gas and groceries has a greater impact on lower income/lower FICO score borrowers, you’d think consumer credit losses would be isolated to lower FICO score bands. But guess what? That’s not what we’re experiencing. Losses are up across the board in our consumer loan portfolio.
COVID-19’s Impact on FICO Scores
During the primary pandemic years (2020 and 2021), the average FICO score rose from 702 to 714, a total of 12 points, or the same amount scores rose from 2012 to 2017, a period when the economy was just starting to recover from the Great Recession. In the industry, it’s believed that the rapid paydown of credit card balances from a combination of COVID-19 stimulus funds and a lack of personal travel contributed to the rise in scores.
So what, you ask? With credit card balances and delinquency rising so quickly in 2023, I’d argue the increase in FICO scores was a blip, unsustainable and a partial explanation for our rise in losses. Humor me and let me share some of my back-of-the-napkin math with you.
Consider a block of what Ent would consider B loans with a FICO score of 660 to 690 that we granted in 2022. We already know that FICO scores increased 12 points for the average consumer. As a result, that block of 660 to 690 loans probably would have scored between 650 and 680 pre-COVID, perhaps even lower. Think about the financial habits during COVID of all consumers with scores from the highest to the lowest. If you believe scores increased due to paying down debt, who do you think benefitted the most? Do you think a block of your members saw their score increase from 820 to 832 because of the credit card debt they paid off? Probably not.
In addition, if you had a block of members who had a 500 FICO score pre-COVID, do you think their score increased to 512 because they used stimulus funds to pay down debt? Again, not likely. Without all the data in front of me, I’d say those consumers who benefitted the most were probably concentrated in that 580 to 680 range. And their scores were likely to have increased by more than 12 points on average since consumers of each end of the score range might not have had a lot of movement.
The real impact? Your B loans originated in 2021 and 2022 (my best guess) are not going to perform to past expectations. You priced the loans like a B, but the pool of loans is going to perform more like a C or C+ because that’s what they really are. Somewhat like the old saying of “buy low and sell high,” you “lent high” and will “collect low.”
Other Trickle-Down Impacts From Score Migration, or Is That Score ‘Inflation’?
Ent CU has been trying to add more B, C and D loans to its consumer loan portfolio for a long time. The Great Recession got in the way from 2008-2013. We ramped up again in 2014, got even more serious in 2018, and had a minor setback with COVID-19 before ramping back up in 2022.
Adding B, C and D loans in a way that avoids near-certain pain and suffering when losses rise is much easier said than done. From 30 years of experience in trying to do this and get the timing right to ensure profitability and success (an article of its own), I can say lending industry executives are like sheep. We’re either trying to take on more risk or trying to reduce risk in our portfolios at virtually the same time. There are a lot of implications of this (another article), but the most important one? When trying to take on more risk, competition for volume tends to drive down rates, ultimately underpricing the risk of B through D loans to build volume.
While taking on more risk without getting paid for it isn’t one of my favorite things, it’s the pursuit of the volume that’s the real problem here. If you think in general terms about B, C and D credit tiers, if score migration pushes consumers into a higher score tier, there are fewer borrowers in each of those lower tiers and more in the A and A+ tiers. Like my earlier sheep comments, when the competition for B through D loans heats up, there are far more lenders for those loans than borrowers who will repay well enough so the loans as a whole will generate marginal profit over a prime loan.
I love the term double whammy, and here it is for lenders chasing yield through taking on more risk. If your credit union had goals to make more B through D loans in 2021 and 2022, you probably had to make more loans from fewer consumers and fewer applications in those tiers. Fewer D credit consumers, and not all D FICO scores are the same. We might approve as few as 15% of D loans because the borrowers are already over-obligated, or that D score means they’re currently past due on little things like their mortgage or car payment. They’re arguably quite a bit different than a D-score member who was delinquent from the loss of their job, but now have all their credit current (albeit that occurred in the last six months), and they can take on additional debt.
Again, so what, you’re wondering? It should have been tougher to make B through D credit loans in 2021 and 2022, so if you met whatever goals were established, you likely had to dip deeper into the risk pool and make loans that you otherwise would have denied pre-COVID due to the ability to repay and a further evaluation of the factors driving their FICO score. You were unlikely to have approved “good” D loans; you probably approved too many “bad” D loans.
This set of events is somewhat like adverse selection in lending. While the term adverse selection is more often used in the insurance business (picture an insurance company that only underwrites homeowners’ insurance in Florida), for decades I’ve believed it applies to lending.
Adverse selection in lending can happen when trying to chase B-D credit, but also trying to get a little too greedy and charging even more for the risk. “Good” B borrowers, who perhaps are more rate-conscious, will go elsewhere. “Bad” B borrowers might not care what the rate is and are only concerned with getting approved. As I noted previously, score inflation as seen during COVID-19 likely saw the “good” B borrowers move into an A range, leaving the worst of the B borrowers behind for you to lend to. As a result, you’re likely to see greater losses across the board, but most prominently in your lower credit tiers.
If you’re responsible for managing the credit bureau dispute process, you fully know the building crisis in credit washing; if this term is new to you, here’s a great article that explains the challenge. The ease of disputing data in credit reports along with a growing industry in “credit repair,” has created consumers who are repeatedly disputing correct credit reporting data until the dispute falls through the cracks. If the lender doesn’t verify the correctness in 30 days, it’s forever removed from the borrower’s credit report. Ultimately this might be another reason we’ve seen consumer credit score inflation.
As opposed to COVID-funded score inflation, this type of score inflation is not well-intended and much shorter-lived. Imagine a consumer who’s had two or three charge-offs. With sufficient time and persistence, if they’re able to wash their credit file of these records, they could see their score increase by a couple of hundred points. Perhaps a rise from the basement to the penthouse! But will they pay like other penthouse inhabitants? It’s not likely.
What to do about this phenomenon? If you’re reliant on automated underwriting, check your criteria. A FICO score is not enough. Do you have limits on the number of tradelines? A credit-washed bureau may wind up looking much like what we call a “thin file.” Just not a lot of information. Do you require a certain number of open accounts or other data such as past high balances or limits? Again, a credit-washed bureau might not be able to stand up to a slightly higher level of criteria and analysis.
If you tend to do more manual underwriting, be on the lookout for files that seem like the consumer has been trying different efforts to get credit or improve their score. For instance, for years I’ve taught underwriters to look for borrowers with a thin file and such creditors as Capital One or others that specialize in easy credit. Also, a thin file that has a lot of newer, smaller limit accounts might be another sign of a washed credit file.
‘Boost Me,’ Experian!
Every year, more and more consumers are aware of the importance of having good credit, along with the cost of having bad credit. All the credit bureaus have programs where consumers can “self-report” or facilitate reporting of their rent and utility payments. Experian is even advertising its own branded debit card that can “build” a payment history to boost the consumer’s score without the resulting debt.
This kind of “everyone gets a trophy” methodology is seemingly positive for consumers. It opens the door to more lenders willing to take a risk on them. But magically granting more credit to people who haven’t had it or have made mistakes in the past isn’t going to magically allow them to pay on time, either. Going back to my college days, this is like scoring on the curve when a chunk of the class didn’t take the exam. Consumers who manage to pay their rent on time will benefit and get a higher score, but they may not pay as well as their new and improved score suggests. And those consumers who don’t pay their rent probably won’t self-report and will certainly wind up paying every bit as poorly as their score suggests. Translation? Another reason for higher losses.
More From COVID
COVID impacted us in so many ways. Remote work. Isolation from friends and family. Fear and doubt. Rash financial and life decisions. What appears to be another impact is the increasing challenge of talking with members who are delinquent, so we can develop options for repayment. Over the last decade, we’ve gradually seen a greater challenge in reaching our members and having a productive dialogue with them. However, COVID-19 seems to have escalated the number of these conversations as much as it has done for digital banking solutions and services.
I’ve had key members of our staff spend more time looking at recent delinquency and loan losses than we’ve had to do in over a decade. I can safely say we’ve never seen so many borrowers who defaulted on their loan, especially their car loan, and failed to have a conversation with our collectors. They obviously don’t know they have a good chance of working out a reasonable plan to avoid default (which through outreach and marketing efforts we can potentially improve), but the discouraging thing is they just don’t seem to care.
Not Deferred Maintenance
Let’s face it, whether it was Congress or the Fed throwing everything at COVID-19 to prevent a deep, long-lasting recession, or credit unions and other lenders offering generous loan payment deferral programs, there was no shortage of strategies to help consumers and businesses.
Two years ago, the industry was wondering: Was the decline in loan losses experienced during COVID-19 due to these aggressive deferment plans, and would there be a price to pay later, somewhat akin to the price of deferred maintenance of your home and car? My best answer? I don’t think what we’re experiencing is deferral defaults.
Here’s why. I have been looking at various metrics. One of them is the ratio of repossessions to the number of loans in our auto portfolio, pre-COVID, during COVID and post-COVID. I don’t see a dramatic decline in this ratio from 2019 to 2020 that could have been caused by these deferment plans.
Unlike an Old-Fashioned Recession …
Recessions are not neat, well-defined events. Consider the Great Recession. Ent CU experienced increasing losses in 2007 and while the economy technically wasn’t in a recession, we could see the storm brewing. While technically the recession ended in Q4 2009, the average consumer would probably say the recession didn’t end until 2016.
Knowing our static pool loss data, I can tell you that loans made between 2011 and 2013 repaid far worse than loans we made from 2014 through 2019. Coincidentally, our 2021 and 2022 loans are starting to pay like those made in 2011-2013 as well. It’s another reason I think our loan performance now is not a result of deferred maintenance. The primary driver of losses is loans we granted after all loan modification programs ended, and the majority of the stimulus funds had been distributed.
That said, there is an end to a recession. Things get better. The economy gets back to “normal.” It did after the Great Recession—but it took a while. Score inflation, higher than normal losses, losses that don’t seem to make sense on seemingly decent borrowers, difficulty in working with borrowers, credit washing, etc.; all these things look to be around for a while. I don’t think they will go away after we recover from the next recession because they don’t have any basis in the economics of a recession. The result? That book has not been written, but I bet in five years from now, we’ll see:
- Higher borrowing costs that will impact more consumers. Not only from higher losses, but I think the Consumer Financial Protection Bureau’s all-out assault on lenders to protect borrowers from misdeeds in the industry, real and imagined, will only add to the costs and expected losses of granting a loan.
- A greater reliance on application data—and not necessarily credit data. Before FICO scores, lenders relied more on income and stability metrics like time on the job and in the area. FICO scores suggested those were all outdated concepts. I think they’ll make a bit of a comeback.
- A re-emphasis on the basics. Over the last two decades, Ent CU has de-emphasized things like verification of income on consumer loans and has relied more on scores. For the lack of a better description, we wanted to make loans as quickly as we could to the best borrowers with few if any hurdles. Whether you’re matching income from the account history of the borrower or getting that data from their other institutions, I think it’s likely we see a little bit of backtracking on how loans get processed.
- A greater use of third-party data. Over the last decade as we’ve watched the rise of the fintechs, we’ve heard more about third-party data. Let’s face it, it’s easy to lie on a loan application. There’s no FICO score equivalent to validate all the data provided on an application. But what if there were? The fintechs have a head start on us, as they’ve had to essentially create proprietary scores that determine the validity of the loan application using other data sources. Without a relationship with their lenders, it’s easier for consumers to “fudge” their application—especially an online one.
So, all in all, it’s a mixed bag. This “consumer credit recession” will speed up the pace of change in so many aspects of lending but may also play out like “Back to the Future.”
Bill Vogeney is the chief revenue officer and self-professed lending geek at $9.9 billion Ent Credit Union, Colorado Springs.