In a recent column for STRATMOR Group, Rob Chrisman pondered whether the industry should even embrace forecasting in a year as unprecedented as 2020 — one where year-end volume forecasts from the MBA and Fannie Mae differ by a whopping 23%.
When volume predictions are inconsistent — or, worse yet, when economists admit they “don’t know” what’s going to happen next — lenders tend to respond by hunkering down and focusing on the here and now. This is understandable, but it doesn’t set lenders up for success in flexing to inevitable market shifts.
Instead of struggling to make sense of sometimes contradictory industry forecasts, lenders should read their OWN tea leaves. Here’s why.
Taking Action in the Short Term
How many times do mortgage executives ask, “How much are we closing this month?” or “How many of these loans in the pipeline are actually going to close?”
Closed-loan volume drives everything: revenue, hiring needs, pipeline management, warehouse line management. How many of you had warehouse line constraints when refis started getting tossed out like candy at a parade? How many of you lost deals because it took so long to move loans through the pipeline that borrowers took their business elsewhere? I even had one industry contact tell me they couldn’t lock refis because they couldn’t handle the volume. That’s insane!
No one planned for 2020’s volume, but after a few months lenders have begun to learn from what is happening. They are adapting to improve their backend processes and even scaling their marketing to keep pipelines full.
Maintaining Perspective Over the Medium Term
As much as the MBA and GSEs differ when it comes to 2020 year-end forecasts, they are united in their expectation of a significant drop in mortgage production in 2021, estimating declines in the range of $700-$800 billion year-over-year.
Is $700-$800 billion a big drop? Absolutely. Is the sky falling? Absolutely not! This drop would equate to roughly $2.4 trillion in volume for next year — still significantly more than 2019’s $2.18 trillion or 2018’s $1.62 trillion.
Of course, just because overall volume may dip doesn’t mean that your volume will. While the MBA and GSEs can tell us how big of a drop to expect, they can’t tell us how it will be distributed. Instead of making business decisions based on global predictions of volume, wouldn’t you be better served running your own forecasting models?
If your own forecast calls for slower production in 2021, it’s also your data — not industry benchmarks — that will tell you how best to handle related business decisions. Mortgage employees are not leftovers in your refrigerator; the “first in, first out” rule does not necessarily apply.
Instead of blindly cutting the “lowest person on the totem pole,” check your data to see if recently staffed employees are actually outperforming more tenured staff. The last thing you want to do is accidentally lay off the next superstar LOA, processor or underwriter. Even if you paid a premium to get that new underwriter, it’s important to use your data to compare costs per individual to their performance level. If their turn times, file quality and so on are head-and-shoulders better than their peers’, it might not make sense to fire them — even if they’re making more. Quality people cost money, and quality performance is what makes your organization stand out from your competition. Assess the data and really understand where cuts make sense.
Planning for the Long Term
Some say 2021 refinance volume will fall far below historic averages, which will create heated competition for purchase loans.
Are you tracking your purchase volume compared to previous periods or years? If you shifted significant resources to focus on refis, did your purchase business suffer? I had an old boss who used say, “Just give me one more refi boom and I’ll be better with the money this time.” Why wait for the boom when you can build a sustainable business now? I get snagging up market share while it’s there, but you can’t lose sight of the bigger picture. Use your data to track whose purchase business is slipping at the individual, branch and regional level and get them refocused on the long game.
If you have a consumer-direct channel, what percentage of that channel is made up of purchase loans? What is their pull-through rate of your purchase business leads? Using data, you can see how often you are converting on those opportunities and where processes could be improved. For instance, you may need to look at your lead aggregators and marketing spend to ensure messaging speaks not only to your ability to deliver the lowest rate, but also to deliver a superb homebuying experience. Use KPIs to track turn times and identify individuals with high fall-out who could fund more loans with a little coaching.
There’s a Better Way
Too many times I hear clients say, “When my ops manager sends me their forecast, I cut it by 25% and see how close we are” or “I use our standard pull-through rate and just take that percentage against the pipeline.” These aren’t forecasts — these are unscientific approximations.
So how do you create predictions that are meaningful? I can tell you what we are doing at LBA Ware. We have built machine learning into our new business intelligence platform, LimeGear. Unlike other BI tools, LimeGear is built for the mortgage industry by mortgage industry professionals. As a result, our machine learning methodology focuses on key metrics specific to the mortgage industry to predict things like what loans in your pipeline are likely to close in the next 90 days (we call this “propensity to fund in 90 days,” or PTF90 for short). The second piece of our machine learning is solely focused on forecasting utilizing historical closed loan data metrics and attributing those key data points to your current pipeline to forecast what would actually close over the next 30 days. This metric can be tied to your warehouse line management so you don’t find yourself scrambling later to find extra money or offload loans at a lesser margin.
What do I do now?
Shameless plug would be to tell you that you should schedule a demo of LimeGear and start building your data strategy for 2021 and beyond. The reality is that data can tell you a lot, but it needs to be presented in a consistent format to have true meaning and allow you to track how your operational changes actually improve overall processes or not. Creating random ad hoc reports is not a data strategy. That is just grasping at straws and being reactive rather than proactive. LimeGear will provide that consistent data strategy so you can drill down and analyze rather than getting lost in the rabbit hole.