where will mortgage volume come from in 2027

If you're building a 2027 plan around a refi wave that's finally coming, stop. The Mortgage Bankers Association's October 11 forecast says the opposite, and the number that matters for your branch isn't the rate path, it's the split between refi and purchase. A flat year means you don't find new volume, you close more of what you already have.

Quick answer: Mortgage volume in 2027 comes from purchase deals and the past-client database you already own, not a refinance rebound. The MBA's October 11 forecast projects refinance originations falling 17% to $571 billion while purchase originations rise just 0.8% to $1.42 trillion, with total loan count near 5 million. A branch owner wins a flat year by closing more of the pipeline, referrals, and dormant database already in hand.

What the MBA 2027 forecast actually says

The MBA projects total single-family origination volume falling about 5% to just under $2 trillion in 2027, with loan count dropping 5.6% to roughly 5.0 million loans. Refinances fall 17% to $571 billion. Purchase originations climb 0.8% to $1.42 trillion. The MBA expects 30-year rates to stay above 7% through 2027, which is why the refi number keeps sliding, per HousingWire's coverage of the forecast.

The practical read: purchase carries the market. MBA's Joel Kan noted refinances account for slightly under 30% of loans, leaving purchases to do most of the work next year. If you planned 2026 around rates dropping and a recapture wave lifting your numbers, that plan is now a bet against the people who make the forecast.

It's worth knowing the houses disagree on how bad refi gets. Fannie Mae's October forecast is more bearish, cutting 2027 refinance volume to about $430 billion while keeping purchase near $1.43 trillion, per Scotsman Guide. Both agree on the shape: purchase steady, refi weak. Plan against the shape, not the decimal.

Where will loan volume come from in 2027

Loan volume in 2027 comes from three places you already touch: the purchase deals in your pipeline right now, the referral partners who send you buyers, and the past-client database sitting in your CRM doing nothing. National Mortgage Professional framed the forecast as more pressure on loan officers to win market share beyond rate-driven refis, then left the hard question unanswered: where will the additional loans come from? There is no new wave. The answer is the deals and relationships you've already paid to acquire.

That reframes the whole year. When new inbound is scarce and purchased leads cost more, the math moves toward conversion over acquisition. One extra closed deal a month from the book you already own beats chasing a market that isn't growing.

Close more of the purchase pipeline you already have

Most lost purchase deals aren't lost to a better lender. They go cold because a human had to remember a follow-up and didn't. Purchase business depends on relationships with buyers, agents, and builders, so the deals that slip are relationship deals, not price deals. A buyer who went quiet after your second text isn't gone; nobody worked them a third time.

Two leaks eat purchase volume all year. First, automated sequences stop the moment a lead replies, so an engaged buyer drops into a human's memory and often falls out of it. We wrote about why follow-up automation dies the moment a lead replies if you want the mechanics. Second, roughly 40% of mortgage leads never get a first contact at all, which we broke down in 40% of mortgage leads never get contacted. In a 0.8% purchase-growth year, those two leaks are the whole game.

Reactivate the past-client database you already own

A loan officer with a few years in the business is sitting on hundreds or thousands of past clients and prospects. Most of them get nothing. Reactivation converts far higher than bought leads: database campaigns typically see 4-10% response on pre-qualification lists and 6-15% conversion on past-borrower outreach, and many mortgage companies hold $1M-$10M+ in recoverable volume, per Prestyj's 2026 benchmarks. That volume exists whether or not refis come back, because past clients refinance, buy again, pull a HELOC, or refer someone.

The reason this stays theoretical for most teams is capacity. UWM's Jonathan Bressler put it plainly in National Mortgage Professional: when a trigger flags 50 or more past clients, there is no shot one loan officer calls them all. The vendors sell the alert. The outreach still needs a system. We covered the recapture version of this in Refinance Recapture: Catch Rate-Drop Refis Before a Competitor, and it matters now because AI-powered digital lenders that sit on servicing data see the payoff request first and recapture your own past client before you know the opportunity exists.

Keep referral partners sending deals

In a purchase-led year, the referring agent is your lead source, and you keep them by never making them call for a status update. The most common complaint on both sides of a deal is the same: the buyer's loan officer went quiet, stopped sending file updates, and the agent stopped referring. Top-producing loan officers treat this as a habit, sending a weekly status update before the agent ever asks. The problem is doing it across every active file, every week, by hand.

Manual follow-up vs a system that works it for you

The gap between knowing the move and making it is capacity. A team of a handful of people cannot work a book of thousands, update every referral partner weekly, and chase every stalled document while still closing deals. Here's the honest comparison for a branch running flat volume in 2027.

This is where SalesOS fits a flat-volume year. It works the follow-up and partner updates your team can't get to, keeps the CRM and your loan origination system (LOS) in sync so nobody retypes a status, and shows where every deal stands without a status meeting. It runs inside the Claude or ChatGPT account you already pay for, and it acts on the private, reversible work on its own while holding anything that goes to a borrower, a partner, or a new contact for your yes. It doesn't replace the relationship selling; it covers the boring, remember-to-do-it work that leaks deals when people get busy.

Common mistakes building a 2027 plan

The forecast changes the plan, but a few predictable mistakes waste the year anyway.

If you're weighing whether to solve the capacity problem with a hire, another tool, or a built system, we laid out the actual numbers in what AI sales automation actually costs a small team.

FAQ

Where will mortgage volume come from in 2027?

Mortgage volume in 2027 comes mostly from purchase deals and your own past-client database, not refinances. The MBA projects purchase up 0.8% to $1.42 trillion and refis down 17% to $571 billion, so growth has to come from closing more of the pipeline, referrals, and dormant database a team already owns.

How is the MBA 2027 forecast different from Fannie Mae's?

The MBA and Fannie Mae agree purchase stays roughly flat near $1.4 trillion but disagree on refis. The MBA projects 2027 refinance volume at $571 billion, while Fannie Mae's October forecast cuts it to about $430 billion. Both point to the same move: plan for weak refi and lead with purchase and database.

Is it worth working a past-client database when rates are above 7%?

Working the past-client database is worth it regardless of rates, because past clients also buy again, pull HELOCs, and refer. Reactivation campaigns typically see 4-10% response on pre-qualification lists and 6-15% conversion on past-borrower outreach, and it converts far higher than bought leads, so it's the cheapest volume a branch can win in a flat year.

How does a small team work a database of thousands when nobody has time?

A small team works a large database by putting a system on the repetitive touches instead of adding people. SalesOS works the follow-up in the team's own templates, keeps going after a lead replies, and sends referral-partner updates automatically, while loan officers handle the conversations that need a person. The team decides the rules; the system does the remembering.

If 2027 is going to be won by closing more of what you already have, the first thing worth seeing is where your own process leaks today. Book a demo and we'll walk your pipeline, your database, and your partner updates to find the deals already sitting there.