
MBA Premier Member Editorial: Tech and AI Are Advancing, So What’s Next for Mortgage Pricing?

Mike Vough is head of corporate strategy at Optimal Blue. In this role, he focuses on business development to support the company’s integration partner network, as well as corporate development to drive its acquisition strategy. Mike also drives the continued growth of Optimal Blue’s data business, including the company’s data reseller partners.
Getting humans to change is a funny thing. Some are motivated by money. Others, by time savings or clout. But in general, we’re all a little change averse, regardless how much we try to claim otherwise. Many of the world’s problems today are attributable to hesitance or flat-out resistance to break accepted routines. This is, in my opinion, a significant factor behind the mortgage industry’s current approach to pricing loans.
Today, conversations around mortgage pricing tend to focus on AI and modern technology, yet there was a time when everything was handled very differently – before the first product, pricing, and eligibility (PPE) engine was even introduced. Mortgage lenders would receive rate sheets by fax or email. Some would print and actually post them on a wall for their lock desk staff or loan officers. A loan officer would call the lock desk to price out a specific borrower scenario, and an analyst would then price out the loan across the different investor rate sheets – by hand, calculator, or for the really forward-thinking, an Excel spreadsheet.
Mortgage investors – i.e., the parties that would go on to purchase the loans on the secondary market – had to communicate their pricing in such a way that a broad group of secondary analysts across a vast set of mortgage lending shops could actually price a loan and then communicate it to a loan officer or borrower. Combine this with the technical limitations of the time, and the current rate sheet we all know (and sometimes hate) was born.
To give you a full taste of this disjointed process, let’s consider the steps. A lock desk analyst would find the loan amount overlay, add it to the FICO overlay, subtract the LTV adjustment – oh, and said overlay couldn’t exceed 2.5 points for certain refinance programs – and then the analyst would check the guidelines of the investor to make sure they would actually buy the loan. Investors would distribute rate sheets with their pricing structures to lenders/correspondents at varying cadences, with each sheet showing portions of the lender or investor pricing structure to the mortgage public.
Enter: the PPE. This technology completely revolutionized product and pricing by dramatically changing a portion of the manufacturing process, making lock desk and secondary marketing teams much more efficient. For the majority of scenarios, the days of manually calculating pricing were gone. Naturally, the PPE concept was designed around the only source of pricing our industry had ever known – our “frenemy,” the rate sheet. Rate sheets went into the PPE. “Pricing” (rate sheet + adjustments/overlays + margins + eligibility/guidelines) came out. Loans were locked. Technology had moved the industry forward.
Of note, even with this big jump in technology and the process improvements that came with it, a lender’s “intellectual property” continued to be managed outside the system, staying within the confines of a lender’s staff. This fact stands today. There are two parts to the secret sauce recipe for mortgage pricing: the margin (i.e., the cost to originate a specific loan program and origination source) and the base pricing (i.e., pricing pre-adjustments). If a lender is pricing best efforts or just a single source, like an agency cash pricing or a single aggregator pricing, the base pricing piece isn’t applicable. This pricing strategy is fine if you are just getting started, but if you don’t create your own proprietary base pricing, you pass control of your pricing to some other entity that doesn’t necessarily share the same goals or concerns as your organization – which is a separate discussion, unrelated to the pricing system itself.
In this early era, most lenders would calculate the margin, or cost to originate, outside of their PPE, possibly in a hedging system, but most often in an internal Excel file or database. The cost of a mortgage would include factors such as underwriting costs, sales cost, or loan officer compensation, and other unknown factors, such as gain on sale, hedge cost, renegotiations, and extensions.
In the past, the typical strategy to create base pricing was to mirror the universe of potential outlets at the point of loan sale or securitization once the loan was funded. This actually wasn’t very hard when investors published AOT or direct trade workbooks that were easy to replicate in Excel or managed by a hedge provider. But this changed about seven years ago when loan sales moved to bulk bidding. Investors retired their AOT or direct trade workbooks and replaced them with APIs around black box models. This gave investors so much more flexibility and allowed them to pay more for the loans they wanted to originate, and vice versa. This perceived “bulk pickup” is now tracked by lenders and fed into their own base pricing, depending on their starting point (e.g., best efforts rate sheets, TBAs, cash pricing, or others). The blending of these sources has been available functionality for some hedge providers for over a decade, but the move to bulk APIs still muddies the pricing for non-aggregators.
As you can see, the journey of mortgage pricing has been a long one, and in many ways, the accepted approach rightfully looks similar today as it did decades ago. Regardless of how much technology the industry adapts, pricing strategy resides with the lender. So, where does pricing go from here? And more importantly, if the industry embraces a shift away from traditional rate sheets and toward more technology, what benefits could lenders see?
First, I would say the move to bulk pricing for loan sales is a trend to watch on the origination side. Instead of investors publishing rate sheets, could bulk APIs move to the point of origination? Maybe – but first, we would need to see further advancements to provide additional detail for audits, to handle historic pricing, manage worse-case pricing, and more. Throw in the growing share of non-agency production, and the move to lender-owned black box API pricing models may still be years away, albeit a likely end point for the industry. This would put mortgages closer to, say, airline ticket pricing.
Second, the calculation of margins is an area where I suspect we will also see changes in the future. In the case of a PPE that also has access to loan sale and hedge performance data, could we see self-calculated or self-updating margins? I believe the train tracks have already been constructed to facilitate this improvement. For example, if the “Mike” branch extended another loan with no upfront fee collected, and the firm’s hedge cost in turn took a hit, the margin could increase the next day on the product, originator, or branch combination – all automatically to account for the increased hedge cost. Why shouldn’t poor operational behavior that directly leads to worse pricing and less profitability stand to be adjusted automatically? Given the advancements we have seen in AI and ML, this isn’t necessarily too far away for our industry – for those who would be willing to embrace it, of course.
So, what’s holding our industry back from such pricing evolutions? Lenders and investors need to embrace incremental tech advancements as they become available. In fact, innovations already exist that are setting us on this path toward more streamlined and profitable pricing practices, such as buy-side loan and servicing APIs, base pricing generation, MSR cashflow engines, and hedge cost and originator scorecard reporting. With the onset of AI, the feedback loop between origination and secondary is rapidly becoming a well-oiled machine, empowering lenders to assess their data and identify trends more quickly in support of more profitable decisions. Closing the timing gap between the primary and secondary markets will allow lenders to more dynamically price. The infrastructure is built, and the possibilities afforded to lenders via this feedback loop will only continue to grow and become more autonomous.
Whatever you do, don’t let the familiar comfort of a rate sheet or an attachment to past processes limit your business. Your “intellectual property” – i.e., your margin and your base pricing – will continue to be managed by you, outside your tech stack. The key to long-term success and market adaptability will be embracing incremental AI and tech advancements as they continue becoming available. The time is now to shift your thinking and prepare for a stronger pricing future characterized by more automation and agility.
(Views expressed in this article do not necessarily reflect policies of the Mortgage Bankers Association, nor do they connote an MBA endorsement of a specific company, product or service. MBA NewsLink welcomes submissions from member firms. Inquiries can be sent to Editor Michael Tucker or Editorial Manager Anneliese Mahoney.)