Thoughts on the Market
2026-08-11 · Hosted by Mike Wilson · Morgan Stanley
Executive Summary
Morgan Stanley analysts Jeff Adelson and Jay Bacow discussed how AI agents could transform the $14 trillion US mortgage market by automating the historically manual process of shopping for and refinancing loans. Bacow explained that only about 30% of borrowers who could lower their rate by 100 basis points (1 percentage point) actually refinance in a given year, largely because comparison shopping is slow and paper-heavy — a gap AI agents could close by instantly querying dozens of lenders and presenting the best option.
Key Stories & Changes
1. AI Agents Could Sharply Increase Mortgage Refinancing Rates
Historically, only ~30% of borrowers eligible for a 100-basis-point rate reduction actually refinance in a given year
Fewer than half of borrowers currently get quotes from more than one lender
An AI agent could query 30 lenders simultaneously, upload documents, and let a homeowner "click a button" to refinance
Bacow's team projects a 40% pickup in refinance volumes versus current expectations in the event of a 100-basis-point rate rally
2. Impact on Mortgage-Backed Securities (MBS) Investors
Faster borrower response makes mortgages more negatively convex (their price behaves less favorably for investors as rates move), shortening effective durations
Morgan Stanley's base case: mortgage spreads widen by roughly 10 basis points as investors demand more compensation for the "prepayment option" they are effectively short
Investors would need to buy more duration to offset shortening, and demand for low-strike rate "receivers" (options that gain value as rates fall) would likely rise
3. Origination Volumes and Lender Economics
Current annual mortgage originations sit around $2 trillion, below a "normalized" estimate of $2.5 trillion
Adelson's bull case sees originations reaching $3 trillion or more — still below the COVID-era peak of roughly $4 trillion
AI could let scaled lenders process more loans with the same headcount, reducing the historical hire-up/layoff cycle tied to refinance waves
However, easier comparison shopping could intensify competition and compress gain-on-sale margins, meaning earnings benefits may lag volume growth
4. Broader Effects: Homeownership, Home Equity, Debt Consolidation
Easier access to multiple lender quotes could modestly expand the homeownership rate by improving borrowers' ability to secure financing
Lower rates combined with AI-enabled shopping could boost use of second liens, HELOCs, and cash-out refinancing
AI agents could also prompt broader debt consolidation (credit cards, auto loans, student loans, mortgage), freeing up consumer spending elsewhere
Analysts expect these effects to materialize mostly on a three-to-five-year timeline
5. Early Signals to Watch
Improving ease-of-use and adoption in mortgage-shopping apps and websites, mirroring broader consumer AI/LLM adoption trends
Faster closing speeds: the industry average is 40-45 days, while top AI-invested originators already close in 12-20 days
An eventual step-up in refinance volumes at current rate levels, as the ultimate outcome of these shifts
Trends Identified
1. AI Is Poised to Compress Mortgage Shopping Friction
The core thesis is that AI agents remove the labor-intensive parts of mortgage shopping — comparing lenders, gathering documents, submitting applications — which historically discouraged many eligible borrowers from refinancing even when it was financially beneficial.
2. Winners and Losers Diverge Between Volume and Margin
While AI-driven origination growth benefits the mortgage industry in aggregate, scaled lenders with strong technology investment stand to gain operating leverage, while increased price transparency from AI-driven comparison shopping could squeeze margins industry-wide — a classic volume-versus-margin tradeoff.
3. Fixed-Income Markets Must Reprice Prepayment Risk
As borrower behavior becomes more rate-sensitive and responsive, mortgage-backed securities investors face a genuine repricing of interest-rate risk, with implications for MBS spreads, duration positioning, and demand for rate-hedging instruments. ---
Sentiment Analysis
Overall Market Sentiment: Constructive, Long-Term
Both analysts framed AI's impact on the mortgage market as a structurally positive, multi-year transformation rather than a near-term catalyst, with clear-eyed acknowledgment of trade-offs for lenders and MBS investors.
Risk Factors Highlighted
Model risk from AI-driven behavior change: Past prepayment models may understate future refinancing speed as AI adoption grows, per Bacow.
Margin compression for mortgage lenders: Increased price transparency from AI comparison-shopping could intensify competition and pressure gain-on-sale margins.
MBS duration and convexity risk: Faster refinancing makes mortgage securities more negatively convex, complicating investor hedging.
Uncertain adoption timeline: Analysts stressed effects are expected over three to five years, with real uncertainty about pace.
This episode was covered in today's [The Market Signal — 2026-08-11](https://marketsignal.beehiiv.com/p/the-market-signal-2026-08-11), a cross-source synthesis of multiple podcast reports.