Part I: Support your collections team now!
In July of this year, I wrote an article in response to several hit-pieces that were published related to subprime auto. That article may be found at the following link: https://nonprimetimes.com/i-hate-to-burst-your-bubble/. My objective was first to debunk talk of an auto loan bubble, and second, to redirect attention to the issues that pose the greatest risk to portfolio performance going into a potential recession. As we conclude 2022, many lenders have already seen elevated delinquency and credit losses, and are bracing for more in the coming year.
There are three primary reasons for the deterioration in performance. First, the substantial forbearance offered to consumers during the pandemic has abated. Since May 2020, lenders saw delinquency drop as much as twenty percent due to pushed off mortgage, rent, student loan, credit card and auto payments. Second, historical trends always show the worst performance in the fourth quarter. Finally, inflation is leading to increased defaults for consumers who are the most debt-stressed, particularly in the riskiest credit tiers.
When new loans are originated, it takes at least six months before data-savvy lenders can see a statistically meaningful difference in early-stage delinquency rates. It takes nearly eighteen months before sufficient defaults accumulate to the point where managers know performance is off track. This is part of the reason companies are frequently slow to respond to shifts in performance. By the time the problem is obvious, there could be two years of toxic paper in the pipeline that has yet to flow through.
This often results in a scenario where servicing executives get blamed for the performance drop-off, and knee-jerk reactions are taken to push an unreasonable collections goal. In some cases, senior collections managers become the scapegoat and get replaced – creating more disruption in a time where the operation needs focus and stability. The overwhelming consensus of economists, industry analysts and leading portfolio managers is that 2023 will result in worse performance due to economic pressures. It is imperative that executive management takes actions that will lead to managing these pressures to a successful and profitable outcome. This begins with properly supporting the servicing operation.
Supporting Effective Collections
When servicing managers are saddled with unreasonable expectations, or blamed for things that are outside of their control, the resulting cure is often worse than the disease. In my career, I have seen unbelievable things done by middle managers who feared for their jobs from overly-aggressive targets. This includes incenting voluntary repos, declaring accounts to be bankrupt that have never filed and paying repossession agents to hide recovered units because they were over the budget for repos allotted by a myopic risk management forecast.
In general, auto finance credit performance is worsening, although it remains within the typical range that is observed across the credit cycle. While there are many opportunities for lenders in 2023 as inventory returns and vehicle pricing drops, they are only available to lenders who aren’t struggling with performance issues. Successfully navigating 2023 is highly contingent on fostering an environment where the collections operation has the best chance for success. This requires that executives consider the following:
• About Face – It is appropriate to analyze collections effectiveness when performance goes south; however, in my experience across many market and economic environments, at least seventy percent of credit quality is determined by those things that take place at the point of origination. Lenders must do an about face and look to originations quality for answers, which must include an analytic deep dive on early stage delinquency at the vintage level using everything known at the point of sale. Identify which attributes are associated with negative outcomes and modify your credit program to improve origination quality. We have performed this type of analysis for many of our clients, and we typically find that close to one third of delinquency and loss comes from around ten percent of originations. When employing this analysis it is not only important to review credit and loan structure attributes, but also shifts in origination behavior that contribute to worse performance. Slower funding time, increased exceptions and poor closure efficiencies can all have an impact that isn’t accounted for in credit data alone. In addition, you may find specific dealers and/or regions consistently represent a higher risk which is not immediately explained purely by credit data. In these cases, it is appropriate to raise minimum score and structure standards provided it is applied in a broad enough way so as not to create a disparate impact issue. Keep in mind that trimming off the worst performance does not necessarily mean a reduction in volume. With improved performance comes the ability to be more aggressive on better quality deals. Collectors must play the hand they are dealt, so when performance shifts the wrong direction the most constructive thing to do is give them better cards to play with.
• Measure Twice Cut Once – This old English proverb simply means if you don’t take your time and measure (or properly assess your situation), you will have to redo your work, thus wasting time and materials. The Russian version says to measure seven times, which I like even better. In the case of collections, there are a finite number of collectors on hand and only so many hours in the day to make calls. Inefficiency in the collections strategy wastes operating expense and leaves dollars uncollected. Collections prioritization models, otherwise known as behavioral scores, take all of the information from origination time, updated credit reports and collections activity over the life of the loan and correlate that to a probability of either rolling to the next delinquency stage or curing. When these models are constructed and deployed properly, they reduce a tremendous amount of information down to a single score that gives the manager a powerful ability to load balance effort to collect more dollars with less expense. For example, a typical roll rate for mid-tier subprime auto paper is twenty percent in the one payment past due bucket. We routinely see results where high scoring accounts have less than a ten percent chance of roll, whereas low scoring ones have more than a seventy percent chance of roll. Consequently, the high scoring accounts that paid typically required less than two outbound call attempts, while the low scoring ones required more than fifty attempts before a payment was made. Knowledge of these probabilities allows for far more productive dialer campaigns and helps to deploy collectors to accounts where they are needed most. These scores are also very useful in overcoming short-term staffing issues. While many large lenders use these tools very effectively, many collections leaders resist them. Their reasons for this are typically related to the inability to move beyond entrenched legacy strategies (i.e., focusing purely on balance and days past due or clinging to a cradle-to-grave method) or fear of taking the blame for a poorly constructed model. The latter issue would not exist if there weren’t plenty of bad models. It is imperative when adopting a behavior score for the first time that the models are built with deep involvement of people who understand servicing operations and that the tool is deployed with a champion/challenger strategy to minimize the fallout from shortsighted models. Accomplishing this requires a strong executive champion to ensure that the analytics team embraces operational input, and that collections leaders know they are fully supported in the rollout.
• The Rule of Reasonableness – The first time I heard of this was in relation to non-compete agreements. For example, in Texas, a non-compete agreement might be deemed unenforceable if the employee was pressured to sign the agreement and no compensation was offered during the period of non-competition. In general, it refers to a legal concept where judges evaluate whether something is done in a responsible or unreasonable way, which can strongly influence the outcome of the case. I am borrowing this term in hopes that executives will apply it to goals imposed on the collections team. Several years ago, I was working with a long-time client to improve their loss forecasting methodology. Their existing risk forecast targeted a monthly net credit loss figure of 0.3% of the outstanding portfolio balance, equating to a 3.6% annualized net charge-off rate. This was ridiculous for two reasons. First, the company’s publicly disclosed performance demonstrated an annualized net loss rate of over 7.5%, which is more than double the risk forecast. Second, the company averaged close to a 600 credit score for originations. That quality paper does not yield near-prime loss numbers, no matter how clever the front-end scorecards may be. There is a reasonable range for delinquency and charge-off by credit tier, an example of which is shown in the table titled Benchmarks for Annualized Net Credit Loss by Niche.
The benchmark illustration should be taken as generic. There are many factors that can influence the range and average, including whether or not the lender has multiple generations of a refined originations score. In addition, the size of the lender, strength of dealer relationships, collections strategy and collateral mix can all have an impact. The important takeaway is that lenders should engage outside parties to help them set reasonable industry benchmarks for delinquency, roll rate, charge-off rate and vehicle recoveries in order to make certain they are not creating an impossible situation for collections managers. Unrealistic forecasts lead to insufficient staffing and unnecessary performance deterioration.
Looking Ahead
The coming year is likely to present a number of challenges to auto lenders. Inflation, supply-chain issues and an increasing debt strain on consumers will undoubtedly become a growing concern. Delinquency issues, which have a way of sneaking up on executives, can reduce the borrowing base, impede funding and put the company at risk with capital partners. I am not suggesting that credit performance in 2023 will be wildly outside the expected range at this point in the cycle, but for lenders who find themselves under pressure, the wrong reaction could exacerbate their problems. Executives who get in front of these issues now, and give the collections team the best chance at success, will be well-positioned to navigate risks in the coming year. The next installment in this series will address declining vehicle values and what lenders can do to mitigate loss.