9.1.26 Fun With Numbers; Zillow’s Meg McGrath Vaccaro on Origination at Search; MSR Historicals

Robbie Chrisman Welcome to the Chrisman Commentary, Daily Mortgage News Podcast. I'm your host, Robbie Chrisman. Topics on today's episode include some fun with numbers, a useful note on the driver of servicing values, and my interview with Zillow's Meg McGrath Vaccaro on Zillow Home Loans' AI strategy from helping buyers and automating mortgage workflows to enabling smarter pricing conversations, and how Zillow's unique ecosystem could reshape the entire mortgage industry. Here, take a listen to a little preview. Why is Zillow positioned differently from a standalone mortgage tech provider or general purpose AI tool? And maybe that's a little reductive of a question because everybody knows just how much of an advantage Zillow has in terms of like two-thirds of Americans using Zillow when they start their home shopping journey. So you have you have that baked in, but but uh the floor is yours. Meg McGrath Vaccaro I think that's exactly the sort of lens to look at this. We are uniquely positioned in the mortgage industry because we are a direct lender that's sitting inside Zillow. So point solutions are really effective at solving a component part of the mortgage manufacturing process. And we certainly leverage the best in class that is available to us throughout that process. But we have a unique opportunity that wouldn't make sense for an AI provider to go after and try to sell to the market because basically Zillow is the only one who has the access to that opportunity. We have the ability to leverage the context window across search and browsing behavior, engagement with real estate, understanding of the mortgage pricing market, and connect all of that to see how we can best guide our consumer to find homes that meet their needs in their budget that they can win and that they will receive financing for. So we are connecting a lot of the capabilities that are available throughout the mortgage industry, but doing it in the way that only Zillow can. And a lot of that necessitates building novel and bespoke solutions. Robbie Chrisman Thanks to Zillow Home Loans, Zillow's in-house mortgage lender, for sponsoring this week's podcasts. By integrating Zillow's real estate platform with financing, Zillow Home Loans helps buyers move from dreaming about a home to holding the keys. With tools built for modern lending, Zillow Home Loans loan officers can focus on guiding buyers with care and confidence. We're now two-thirds of the way through the third quarter of 2026, and lenders and vendors are adapting to a lackluster home buying season. Stubborn rates and origination costs around eleven thousand dollars per loan. Lenders are trying to drive that cost down through higher pull-through. Labor Day is next Monday, and talk like a pirate day is the 19th. Loan originators are watching the demographics of people in their twenties and are also following trends in the rental markets and with landlords. For example, in 2022, more than half of American households were childless. 29% were married households without children, and 28.9% were single households without kids. More than a quarter of the households included parents. 17.8% were married households, while 8.1% were single parent households. Certainly, households are changing. Does your product mix, originator base, and investor mix reflect that? It seems especially important now as the mortgage industry confronts a combination of compressed margins, elevated operating costs, weak housing supply, difficult economics, and technological change that is forcing companies to decide whether they are built to endure or merely built to grow. I interviewed Bill Cosgrove for an exclusive article on the Chrisman Commentary website about this. I'm only saying this because I think it's highly worth the read. CrispinCommentary.com. It'll be on the homepage. Bill's argument is straightforward but consequential. The industry cannot count on the gross margins of the past returning, so the path forward is to fundamentally change the economics of originating a loan using technology, including AI, and operational discipline, to cap and eventually reduce the cost to close. US Treasury yields moved sharply higher to open the week as renewed US Iran hostilities pushed oil prices higher. The tenure yields surpassed 4.75% for the first time since January of 2025, and the five-year reached its highest level since early 2025. Thirty-year yields broke above last week's highs due to growing concerns about persistent inflationary pressures, geopolitical risk, and the sustainability of elevated U.S. borrowing needs following in the country's debt surpassing $40 trillion. Investors received no notable economic data yesterday, but the rest of the week will bring some noteworthy reports, such as today's August ISM manufacturing index. Markets are now pricing in more than a 60% chance of a September 16th Fed hike after Warsh's hawkish Jackson Hole remarks, with resilient employment and sticky inflation making a hike increasingly likely, unless upcoming payrolls, CPI and PPI data provide a compelling case to hold. From helping buyers and automating mortgage workflows to enabling smarter pricing conversations, and how Zillow's unique ecosystem could reshape the entire mortgage industry. She's senior director of product for Zillow Home Loans at Zillow, where she leads teams building experiences, products, and tools that help simplify the home financing journey and support more efficient, transparent transactions. She has nearly two decades of experience across product management, user experience, and consumer technology. We just met, but I'm going to dive right in here and say how does Zillow Home Loans think about the right role for AI? Meg McGrath Vaccaro Absolutely. So we start from the fundamental belief that home buyers are going to always want to seek human advice and guidance throughout their process, even in the age of AI, with increasing opportunities for automation and technology to play an increasing role in that process. It's a really big consequential decision that people have to make. The customers that ZHL has the opportunity to serve are going to skew a lot more towards first-time home buyers. And it's not uncommon for them to be first generation home buyers. So they're not benefiting from prior experience with getting a mortgage. And they may not even have a vast network of folks to turn to and ask advice for. So they really value getting on the phone and talking with people who they can trust to guide them through this process. So as we think about AI, we uh and technology writ large is we're looking for opportunities to give our loan officers the most amount of time to invest in those personal relationships, provide that consultation throughout the process and the tools necessary to deliver the personalized service that our customers demand. Robbie Chrisman I've always associated Zillow with the very beginning of the home shopping process. Before people even think about getting a real estate agent, before people even think about getting a lender, they're on Zillow, browsing around, probably looking at homes out of their price range that are tantalizing, but it is truly the start of the home shopping process. Now, when it comes to advice and consultation, I think people are turning more toward AI. They can get a lot of answers on whatever platform they use in ways that they couldn't before. Maybe that maybe they felt like originators were gatekeepers in some way, or real estate agents were gatekeepers of information. What does AI change for a buyer at the very beginning of the home shopping process, in your estimation? Meg McGrath Vaccaro Well, Zillow has the opportunity, as you're as you're highlighting, to serve like two-thirds of the American home buyers. And our data says that they are starting that journey, as you're highlighting, months prior to transacting. At the point that we have the opportunity to meet them, they aren't asking questions about like what rates I should select or what kind of loan product should I explore. They're asking really fundamental questions about like, should I dare to dream about home ownership? Could I actually afford this home or this neighborhood? And at that point, they need help asking questions around affordability. And that's why we've built features like BuyAbility, where we can provide more than just an affordability estimate. It's affordability backed by Zillow home loans, real pricing and eligibility. But what we found is that when we answer those types of questions, we introduce new questions. And because these folks don't have that vast network of people in their life to turn to and ask, the alternative would be getting on the phone with a loan officer. But folks at that point in their journey are nervous to take that step. So they're looking towards now, in this moment, asking these questions of AI. But if we lose them to their AI chatbot of choice, we lose the thread in the conversation. So Zillow is investing in building AI mode that actually brings that same kind of capability from ChatGPT or Anthropic or whichever model that you choose to engage with right into the Zillow experience. So that as you're shopping, you're confronting can I afford this home? We're giving you that inline integrated affordability experience plus pairing it with an opportunity to have a conversation, but we make it a digital forward conversation so that we remove all of the worry about being judged by another human or getting locked into a lender selection before that you're really ready to make that kind of choice. The thing that's interesting about how we're approaching that though is because we're building this integrated transaction with deep integration across the modalities of the search experience, the partnership with your real estate agent and Zillow Home Loans, the direct lender, as that conversation moves towards uh integrating with a loan officer, all of the context that the customer has been giving and sharing with Zillow is now available to their loan officers so that we can continue that conversation and pick up where you left off. And they can move seamlessly between the digital experience and our service-powered experience through our loan officers. And that continues not just at the top of the funnel with viability, but even as we progress down funnel with shop with pre-approval. So if the customer gets to a verified pre-approval letter, now they can leverage that same verified pre-approval pricing information as they're browsing Zillow. Robbie Chrisman It's very, very cool to hear. Should they dare to dream of owning a home? Yes. Can they afford that home in that neighborhood? Maybe. We'll see. I'll give them the answers. Let's dive in to this a little bit. I want to talk about how AI is being used within the mortgage application and underwriting workflow today. Just where are we with that? Meg McGrath Vaccaro We said earlier that the most important thing is to create capacity for our loan officers to invest in that relationship building, both with our borrowers and their agents. In order to do that, we have to leverage technology to drive efficiencies in the process. So we're investing heavily in sort of table stakes opportunities for technology in the manufacturing process for originations and fulfillment. So certainly we're doing things with document processing and data extraction calculations, sort of risk assessments. We're starting to do really interesting things with AI agents in the fulfillment space. That's part of the equation. But as we create more capacity, we reduce the manual effort necessary to manufacture, we have the opportunity to do that reinvestment. And when our loan officers are faced with that opportunity to engage home buyers and agents, we want to empower them with insights that really elevate the service that they're able to provide. So we're investing in tools to leverage the data that Zillow uniquely has access to. We have invested heavily over the past many years in generating market insights. So we're pulling that in so that our loan officers who are operating at a national scale can speak to confidently about more local dynamics. As we expand our loan product portfolio, it's going to be increasingly difficult for loan officers to hold in memory all of the potential solutions they have at their fingertips to meet our customers' needs and navigate their eligibility. So we're investing in tools, we're calling it smart scenarios to help aid our loan officers in really capitalizing on the opportunity that Zillow Home Loans is affording them. Robbie Chrisman Yeah, let's talk about SMART scenarios a little bit. Zillow Home Loans is piloting these to support pricing conversations. How do they work? What does it actually change for an LO? Meg McGrath Vaccaro Zillow Home Loans is aggressively expanding the portfolio of loan products. We're going to be going into increasingly local products. And that's going to be hard for a loan officer to remember, you know, a local down payment assistance program in some uh county when they may never see another customer from that county again. So it's going to be challenging to hold that all in memory. So we're investing to make it easier for them to exploit that full opportunity set without having to rely on their own memory and training of that option. So Smart Scenarios sits on top of our proprietary pricing solution inside our loan officer portal. The loan officer gives our customers eligibility information and their goals. Are they optimizing for a specific budget on their monthly payment or their cash to close? And we'll generate that full result set across our portfolio. And then we'll help navigate the decisions that a loan officer has to make to optimize for those things that our customer really is caring heavily about. A lot of customers really think more in terms of monthly payment than overall purchase price. So finding the levers that the loan officer has to pull across that vast portfolio of loan products and really guiding them to make the choice. So the loan officer is still making the choice, but we're helping with the recommendations that land on the optimal solution. Robbie Chrisman To bring this conversation full circle, I want to close by asking you what you think AI will change about the mortgage industry over the next couple of years. We started by saying, how does Zillow Home Loans think about the right role for AI? So, how we're thinking about it now into how is it going to transform the industry over the next couple of years into the future? Meg McGrath Vaccaro Absolutely. I think that it would be foolish to say that we're not going to automate manufacturing. In many ways, I think we have the capabilities necessary to do so now, at least for a meaningful portion of the opportunity that flows through at least our doors. That it's it's going to come down to whether or not we have the time and capabilities to make the investments necessary to do that automation. But customers don't want to push a button and get a mortgage. They want to really engage with the optionality that they have to meet their needs and their goals. And the the key insight is that when they start this journey, they don't really have the information necessary to articulate with specificity about what those goals are. They're going to be dependent on the actual market dynamics that they are experiencing at the time that they are engaging in this process. So they need to be making trade-offs relative to the actual inventory that's available to them. So the manufacturing is an important piece to automate, but only insofar as it helps us reinvest in the guidance that's necessary to meet the moment that our customers are experiencing. As we tackle that harder problem, it's going to position loan officers and other operators in the process in a different role. And we need to support them in understanding what customer questions are coming their way and what data and insights we have to support them and actually rendering that service. When we succeed in meeting that moment and joining that new conversation, I think we have the opportunity to not just better serve home buyers, but create better homeowners that are more prepared for the ongoing responsibilities of owning homes. And that's going to transform how I think people show up in their communities as well. So I'm excited for the more diverse population of home buyers that we're going to have the opportunity to serve once we can start engaging with them on a more personalized way, meeting them at the moment that they're in with uh the level of education that they have on the home buying process and see how we can start transforming who gets to be a homeowner in this country. Robbie Chrisman Yeah, it's a huge opportunity. It's incredible to see Zillow spearheading that. I wish you all the best of luck. Meg, I thoroughly enjoyed this. Hopefully, this is the first of many conversations we have. So thank you. Meg McGrath Vaccaro Thank you so much for having me. Robbie Chrisman Clearly, rates are highly sensitive to incoming data and the Fed's evolving policy signal, particularly after the two-year yield posted its largest one-day jump around a Jackson Hole speech in decades. With only two meetings remaining after September and the June dot plot already showing a modest 2026 hiking bias, upcoming reports that show sticky August inflation and or solid employment could push the new dot plot towards signaling more additional tightening this year, reinforcing upward pressure on front-end yields and increasing the importance of the September decision for the broader rate outlook. Mortgage servicing rights or MSR pricing history shows that mortgage rates are the dominant long-term driver of MSR values, but the same mortgage rate can produce very different MSR valuations depending on Treasury yields, curve shape, mortgage basis, spreads, and liquidity conditions. Accordingly, MSR models calibrated to individual periods explain pricing far better than a single model spanning a longer time horizon. Put another way, valuation relationships themselves change over time. In today's higher for longer environment, bulk MSR multiples appear elevated in absolute terms, but that may be less meaningful given historically high mortgage rates. While new issue, SRP pricing has risen surprisingly little relative to primary rates, creating a noteworthy divergence between the two markets. Today's economic calendar kicks off shortly with final August S&P Global U.S. Manufacturing PMI and will be followed by July job openings, July construction spending, and August ISM Manufacturing Index. Markets will also receive remarks from a Fed Vice Chair for Supervision Bar. We begin the day with agency MBS prices slightly better than yesterday's close, the two-year yielding 4.36, and the 10-year yielding 4.78 after closing yesterday at 4.76%. Let's wrap up with a joke and some housekeeping. A woman goes to a pet store and buys a parrot. Before she leaves, the owner warns her that the parrot had previously lived in a brothel and might have picked up some salty language. When she gets the bird home, he looks around and says, New house. She puts the bird down in the house, and the parrot observes her for a moment before saying, No, madam. The woman shrugs and thinks that it's not so bad. Shortly after the woman's two daughters return from school and see the new pet, the parrot looks at them and says, New girls. Then her husband comes home from work, and the parrot looks over at him and says, Hi, Bill. Thanks again to this week's podcast sponsor, Zillow Home Loans, Zillow's in-house mortgage lender. With tools built for modern mortgage lending, Zillow Home Loans loan officers can focus on guiding buyers with care and confidence. To learn more, visit Zillow.comslash home loans.