The Doom Loop

Why this week's rate move just transferred market share between mortgage originators

Last week, a fifteen-year originator emailed me because he had been losing deals he used to win and could not figure out why. Rates had moved. Borrowers were shopping. Realtors were pressuring him to match. His company had told him they could not sharpen the rate. He wanted to know what was happening to him.

The answer was structural. He had not forgotten how to sell. His company’s cost stack showed up.

This is what I call the doom loop. It is the second-order consequence of an inflated cost stack in a rising-rate environment. It is happening right now in every mortgage shop that has not rebuilt its operations. And it is the structural reason the gap between efficient and inefficient mortgage companies just widened by an enormous amount in the last seven days.

The Spine of It

A company with a heavy cost stack cannot price competitively. To survive, it raises margin to protect P&L. Higher margin makes its originators less competitive, which lowers volume, which raises cost per loan, which forces the company to raise margin again.

The people inside it usually blame themselves. The originator looks at the lost deals and wonders if he is slipping. The branch manager looks at the P&L and wonders if she is managing tightly enough. Neither sees the cost stack sitting upstream, unmoved by any of it. The structure was engineered so they wouldn’t.

How the Machine Works

Most mortgage companies are structured the same way regardless of size. Corporate takes its margin first. The MBA reported corporate overhead averaged $5,785 per loan in 2025. This is the fixed company scrape. It does not move when margins compress because it is mostly fixed costs and labor.

The originator takes their commission second. The MBA reported average originator compensation of $3,777 per loan, or 30% of the average $12,500 in gross revenue.

The branch gets what’s left.

Under normal 300 bps gross margin conditions this works. The lock margin in our pipeline is 235 bps as I write this on May 21. A month ago it was 290 bps. Rates went up. Margins compressed. Now the structure really matters.

In most structures, when a borrower asks for a rate concession, the concession does not come out of corporate’s $5,785. It does not come out of the originator’s $3,777. The concession comes out of the branch.

A friend at another company who used to run a P&L branch described to me what that looks like. He got a $17,000 concession on a $2 million jumbo approved during a slow winter. His company calculated branch margin on a 90-day rolling average, so the single concession rewrote his calculated margin on every loan that closed for the next ninety days. This is the dumbest thing in the world. The jumbo still brought in $20,000 in revenue. By the time he understood what had happened, the branch was at negative $14,000 on the loan because corporate’s margin demand was more than what was left.

To make up the loss, he raised his branch margins. His originators couldn’t compete. They quit. Then he quit too.

All over a loan that still brought in $20,000 of revenue.

Another originator I know in a corporate model, doing about $120M a year for the last several years, said it differently. He told me, looking at his own numbers on the screen, “I’m almost incentivized to drive profits to zero. There being money in it doesn’t benefit me.” The structure he was in had trained him, year by year, against caring about operating well. Giving concessions was the only way he benefited. And then last week, his company told him he could not give one.

That was the email I opened this newsletter with.

What the Rate Spike Did

Then rates moved.

Borrowers shopped. Realtors pressured. Every unlocked deal turned into a rate fight.

Corporate’s $5,785 did not move. And as volume goes down, the cost per loan goes up.

The originator’s commission did not move. The branch absorbed all of it.

So the branch did the only thing the structure allowed. It raised margin.

The originator who was at par 6.5% last week showed up Monday at 6.625% or 6.75%. He filed PE requests. The divisional director slow-walked them. The deals slipped. Volume dropped. Cost per loan rose. Margin went up again.

This is the doom loop in real time.

I ran this exact structure at Princeton for about eighteen months after I exited the wholesale channel in 2022. The math did not work then. It still does not.

What This Looks Like in the Market

Last month, a borrower called one of our new originators with a loan estimate in hand.

The LE was from the lender his realtor referred him to. A lender who aggressively advertises low rates on social media.

• Rate: 6.25%

• Points charged: 1.25% ($5,891)

• Application fee: $1,550

• Total origination charges: $7,441

• Lender credit: $0

• Total closing costs to borrower: $30,290

• Cash to close: $18,984

• Monthly P&I: $2,901.91

• Estimated revenue to that lender: over $24,600

The second LE was from our originator.

• Rate: 6.00%

• Points charged: 0

• Total origination charges: $1,595, credited back to $0

• Lender credit: $1,595

• Total closing costs to borrower: $22,652

• Cash to close: $11,346

• Monthly P&I: $2,825.72

• Revenue to Princeton: $13,663

Our originator got the loan.

And then he got the realtor as a new referral partner. The realtor did 43 buyside transactions last year. Not bad.

The borrower kept $7,638 at the closing table. He saves $76 per month for 360 months, which is another $27,428 over the life of the loan. Total savings to the borrower for picking up the phone and getting a second LE: roughly $35,000.

Princeton made $13,663 of revenue on the loan. The MBA reported the average mortgage in 2025 generated $12,500 in revenue. We made $1,163 above the industry average on a loan with no points charged. We’ll do this loan for $13,663 of revenue all day long.

The other lender was trying to generate $24,600 in revenue. They got $0.

Pigs get fat. Hogs get slaughtered.

Same loan. Same week. Same market. Different cost stack.

The originator on our side has been in the business two years. He has been at Princeton for two months. His pipeline right now is the largest of his career. I just got off the phone with him to confirm the details of what happened. He said that he is winning deals he would have lost six months ago. He is being referred by realtors who six months ago would not have taken his call. In his exact words: “I’m pumped.”

He is having the best month of his career because rates went up.

Why This Is Possible

In April, our all-in manufacturing cost — opening, underwriting, closing, and post-closing — was $535 per loan. Including processing, we were at $1,003. A well-known industry consultant looked at our cost structure last week and said, “You are a unicorn.”

Jim Deitch’s most recent MBA report puts the industry’s fulfillment cost at $2,454 per loan. That alone is a $1,451 per-loan gap to where we are operating, or roughly one-eighth in rate.

Run the comparison on the whole cost stack above the originator and the picture gets sharper.

• Princeton: $5,000 per loan ($2,000 non-commission sales cost + $3,000 corporate)

• MBA average: $8,785 per loan ($3,000 non-commission sales cost + $5,785 corporate)

• Cost gap: $3,785 per loan

Each one-eighth in rate produces about $1,845 of revenue on the average loan. So a $3,785 cost advantage converts to roughly one-quarter in rate. Or 49% more originator compensation at the market rate (150 bps vs 100 bps). Or some combination of the two.

The originators losing market share right now are not losing because they are worse salespeople. They are carrying $3,785 more in cost per loan than the originators taking their business. In a calm market that gap is a competitive disadvantage. In a rate-spike market it is the difference between winning and losing every shopped deal.

A rate spike is a stress test on the cost structure underneath the rate sheet. Skill is not the variable. Cost stack is the variable. The hard market is when the variable becomes the only thing that matters.

How to Know If You Are Inside It

If your company has told you they would lose money on the deal you are trying to win, you are inside the doom loop.

If you filed a PE request and you are still waiting on it or had to throw a fit, you are inside the doom loop.

This was not theoretical on May 21.

Our conventional locks averaged 6.49%. HousingWire’s Rates Center pegged the national average at 6.88%.

Our FHA locks averaged 6.125%. HousingWire pegged the national average at 6.42%.

That is what cost structure looks like when it reaches the borrower.

If you have lost a deal you would have won six months ago and you have been wondering whether you are slipping, you are inside the doom loop. You are not slipping. The math is structured against you.

The structure is not fixable from inside the company you work for. The cost stack above you was built over years. The people managing it are not going to voluntarily unwind the departments, overrides, approval chains, reports, and meetings that justify it. It is fixable only by moving to a structure where the cost stack is visible, where the bloated sales management layer does not exist, and where you can be competitive on price.

Rates will eventually settle. They always do. The gap between efficient and inefficient cost structures will not. It is widening right now, this week, while you read this newsletter.

The originators who understand which half they are in are taking market share.

The originators who do not are blaming themselves for losing it.

The numbers don’t lie. They don’t negotiate. They don’t care how this week felt.

They simply describe the world as it is.

Go be unleashed.

— Rich Weidel III

CEO, Princeton Mortgage