Why Non-QM Rates Are Higher — and How to Defend the Number

There's a moment on nearly every Non-QM file where the borrower — often a sophisticated, successful, financially literate borrower — looks at the rate and asks why it's a point or two above the number their neighbor got. If you fumble that moment, the file cools. If you handle it well, you get trust for the rest of the transaction. The answer isn't a script. It's understanding what actually builds the number.

Where the base rate comes from

A conventional loan gets sold into a government-sponsored channel with an implicit guarantee behind it. That guarantee makes the paper deeply liquid and cheap to fund. A Non-QM loan doesn't have it. It's bought by private capital — securitization investors, insurance money, funds — who price it against alternative yields and against the fact that the secondary market for it is thinner. When capital markets get nervous, Non-QM spreads widen faster than agency spreads, which is why your Non-QM rate sheets sometimes move on days when the 10-year barely budged. It's a liquidity premium as much as a credit premium, and saying so out loud makes you sound like someone who knows the market.

The adjustment stack

Base rate is only the starting point. Every rate sheet applies adjustments, and the borrower's final number is the sum of them. Understanding the stack is what lets you improve a quote instead of just delivering it.

• Doc type — full doc prices best, then 1099 and P&L, then 12- and 24-month bank statement, then DSCR, then no-ratio.
• LTV — the steepest single lever. Dropping from 80% to 75% often improves pricing more than a 20-point FICO gain.
• FICO band — priced in tiers, usually in 20-point steps. A borrower two points below a tier line should know it.
• Occupancy and property type — investment costs more than primary; condotels, non-warrantable condos, and rural add on.
• Prepayment structure — on DSCR, accepting a longer prepay penalty commonly buys a materially better rate.
• Loan amount — very small loans and very large loans both carry adjustments at the tails.

Move a lever before you shop a lender

LOs new to Non-QM respond to a high quote by calling another investor. Often the better move is restructuring the file. If the borrower is at 80% LTV and one FICO point under a tier, an extra 5% down plus a rapid rescore can change the pricing more than any lender swap. If it's a DSCR investor who plans to hold the property for a decade, accepting a three- or five-year prepayment penalty they were never going to trigger can buy a rate improvement worth thousands a year. If the ratio is at 0.98 and interest-only qualification is allowed, the IO structure can lift the file into a better band entirely. Those are conversations only an LO who understands the stack can lead.

The reframe that actually works

Never let the comparison be Non-QM rate versus agency rate, because that's a comparison against a loan this borrower cannot get. The honest frames are better anyway. For an investor: what does this property earn me at this rate, and is that return acceptable? A property cash-flowing at a Non-QM rate is a property they own, versus a deal they didn't do. For a self-employed buyer: what is the cost of waiting two years to show different tax returns, in a market where the house they want appreciates and their rent doesn't stop? For a credit-event borrower: this is a bridge, priced as a bridge, and the refinance conversation starts in eighteen to twenty-four months once the event seasons and the payment history rebuilds.

Know the sheet, own the room

The LOs who win Non-QM business aren't the ones with the lowest rate — the market is too fragmented for anyone to hold that title consistently. They're the ones who can explain, in ninety seconds, exactly why the number is what it is and what specifically would change it. That confidence is learnable. Pull up your investors' rate sheets, walk the adjustment stack on a real file, and use the Scenario Desk when you need to know how a specific investor prices a specific wrinkle — it answers from their actual published sheets, with the page cited.